microbial risk assessment modeling for exposure to land ... risk assessment modeling for exposure to...
TRANSCRIPT
Microbial Risk Assessment Modeling for Exposure to Land-Applied Class B
Biosolids
A Thesis
Submitted to the Faculty
of
Drexel University
by
Jingjie Teng
in partial fulfillment of the
requirements for the degree
of
Doctor of Philosophy
May 2012
©Copyright 2012
Jingjie Teng. All Rights Reserved.
ii
Acknowledgement
First I would like to thank my supervisor Dr. Patrick L. Gurian, and my co-
supervisor Dr. Mira S. Olson, for their sustained enthusiasm, creative suggestions, and
exemplary guidance throughout the course of my doctoral research at Drexel University.
I feel so lucky that I had two advisors for my graduate studies because I have received
twice of good advice and supports from them.
I would also like to thank my thesis committee members, Dr. Charles N. Haas, Dr.
David E. Breen, and Dr. Sabrina Spatari, for their comments and suggestions on this
thesis.
Third, I would like to thank Dr. Charles P. Gerba, Dr. Ian Pepper, and Dr. Irene
Xagoraraki for their research work and comments on this study.
Forth, I would like to thank my colleague, Dr. Arun Kumar, Dr. Tao Hong,
Heather Galada, Alrica Joe, and Haibo Zhang. This study could not be completed without
their hard work and help in many different ways.
Last but not least, I would like to extend my thanks to my husband Ran Liu and
my friends for always being there for me.
iii
TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................. iv LIST OF FIGURES ............................................................................................................ v CHAPTER 1. INTRODUCTION ....................................................................................... 1
1.1 Background ....................................................................................................... 2
1.2 Literature review ............................................................................................... 4
1.3 Research objectives ......................................................................................... 13
CHAPTER 2. MODEL DEVELOPMENT - ADDRESSING WET-WEATHER EVENTS
........................................................................................................................................... 17 2.1 Introduction ..................................................................................................... 17
2.2 Methods........................................................................................................... 18
2.3 Results ............................................................................................................. 25
2.4 Discussion ....................................................................................................... 35
CHAPTER 3. MODEL DEVELOPMENT - SUBSURFACE FATE AND TRANSPORT
OF PATHOGENS ............................................................................................................. 36 3.1 Introduction ..................................................................................................... 36
3.2 Methods........................................................................................................... 38
3.3 Results ............................................................................................................. 45
3.4 Discussion ....................................................................................................... 52
CHAPTER 4. MODEL INTEGRATION - SPREADSHEET ENVIRONMENT............ 58 4.1 Introduction ..................................................................................................... 58
4.2 Methods........................................................................................................... 59
4.3 Results ............................................................................................................. 64
4.4 Discussion ....................................................................................................... 66
CHAPTER 5. MODEL APPLICATION – UNCERTAINTY ANALYSIS AND
SENSITIVITY ANALYSIS ............................................................................................. 68 5.1 Introduction ..................................................................................................... 68
5.2 Methods........................................................................................................... 69
5.3 Results ............................................................................................................. 75
5.4 Discussion ....................................................................................................... 83
CHAPTER 6. CONCLUSION.......................................................................................... 86 BIBLIOGRAPHY ............................................................................................................. 91 APPENDIX ..................................................................................................................... 103
iv
LIST OF TABLES
1. Table 1-1 Major pathogens potentially present in municipal wastewater and
manure (U.S. EPA 2000) ……………………………………………...………….5
2. Table 1-2 The pathogen reduction standards (Part 503 Rule pathogen density
limits) for Class A and B biosolids (U.S. EPA 2000)…………………………….6
3. Table 2-1 Calculated precipitation values (cm) for Ingham County (Michigan,
U.S.A.) …………………………………………………………………………..21
4. Table 2-2 Site-specific parameters for lab experiments study…………………...26
5. Table 2-3 Comparison between the predicted and observed infiltration depth….27
6. Table 2-4 Predictions of infiltration and runoff values produced in loam soil for
100-year return period rainfall………………………………………………...…33
7. Table 2-5 Critical rainfall values for runoff and infiltration in loam soil for surface
water and groundwater transport scenarios……………………………...……….34
8. Table 3-1 Summary of model validation results………………………………....44
9. Table 3-2 Microbial breakthrough information and indicator:pathogen ratios in L2,
L5 and L6……………………………...…………………………………............56
10. Table 4-1 Site-specific input parameters……………………………...………....66
11. Table 5-1 Site-specific conditions….…………………………...……………….73
12. Table 5-2 Critical rainfall event information (see Chapter 2 for source)………..74
13. Table 5-3 Cumulative risks over time for exposure from five pathways with
uncertainties……………………………...…………………………………........77
14. Table 5-4 Input-output correlations for risks of illness cumulative over time from
adenovirus……………………………...…………………………………...…....80
15. Table 5-5 Ratios of pathogen:indicator in biosolids and in the environment……81
16. Table 5-6 Comparison of enteroviruses risk estimates…………………………..85
17. Table A-1 Compilation of occurrence by pathogen (items in red are data
gaps) ………………………………………………...……………………….....103
18. Table A-2 Microbial decay in ground water (units in 1/ hour)…………………106
19. Table A-3 Microbial partitioning values (Chapter 1)… ……………………….109
20. Table A-4 Dose-response models for different biosolids-associated bacteria.....110
21. Table A-5 Inputs describing site characteristics and application events (Chapter
3)……………………………… …………………………...………………..…112
22. Table A-6 Model inputs with uncertainties: microbial parameters (Chapter 5)..113
v
LIST OF FIGURES
1. Figure 1-1 Literature values for annual infective risks (in log10 scale) from land-applied
biosolids through different exposure routes ……………………………………………..11
2. Figure 2-1 Intensity-duration-frequency relationships for Ingham County (Michigan,
U.S.A.) at different return periods ……...……… …………………………...………….22
3. Figure 2-2 Infiltration (i.e., wetting front depth) and runoff (i.e., surface runoff)
produced by loam soil for a 100-year return period rainfall in relation to rainfall duration
and intensity……………………………...…………………………………...………….29
4. Figure 2-3 Infiltration (i.e., wetting front depth) and runoff (i.e., surface runoff)
produced by different soil texture classes for a 100-year return period rainfall in relation
to rainfall duration and intensity……………………………...…………………...……..31
5. Figure 2-4 Infiltration (i.e., wetting front depth) and runoff (i.e., surface runoff depth)
produced by loam soil in relation to rainfall duration and intensity………………….….32
6. Figure 3-1 Groundwater exposure model……………………………………………..39
7. Figure 3-2 Flow chart for model development……………………………...………...40
8. Figure 3-3 Predicted and observed breakthrough curves for chloride…………….…..46
9. Figure 3-4 Comparison of the fitted velocity and measured velocity (calculated
velocity from field measured infiltration rate divided by porosity) in six lysimeters...…48
10. Figure 3-5 Predicted and observed breakthrough curve for P-22……………………50
11. Figure 3-6 Comparison of predicted concentration breakthrough curve for P-22 and
adenovirus……………………………...…………………………………...…………....55
12. Figure 4-1 Exposure pathways considered……………………………...…………...61
13. Figure 4-2 Flowchart of the SMART Biosolids model……………………………...64
14. Figure 4-3 Columns show representative nominal risks across pathogens for exposure
through groundwater. Error bars represent the 5th
and 95th
percentiles………………….66
15. Figure 5-1 Plots show cumulative risks over time for exposure through five pathways.
Error bars represent the 5th
and 95th
percentiles. Risks for adenovirus, Cryptosporidium,
enteroviruses, and Giardia lamblia are risks of minor illness cumulative over time; risks
for Salmonella and Shigella are risks of major illness cumulative over time…………....76
vi
ABSTRACT
Microbial Risk Assessment Modeling For Exposure To Land-Applied Class B Biosolids
Jingjie Teng
Patrick L. Gurian, Supervisor, Ph.D.
Mira S. Olson, Co-Supervisor, Ph.D.
Biosolids has been used as a soil amendment to enhance agricultural production.
While providing benefits to agriculture, land application of biosolids may introduce
pathogens into the environment and present human health risks. There have been several
studies on the link between land-applied biosolids and human health. However, land-
application sites vary, making it important to have models that can be implemented for a
site-specific assessment of risk.
This study developed and applied a spreadsheet-based tool, named The
Spreadsheet Microbial Assessment of Risk: Tool for Biosolids (SMART Biosolids),
which links quantitative microbial risk assessment with microbial fate and transport
modeling. The SMART Biosolids model estimates risk associated with exposure to
pathogens from land-applied biosolids through five pathways: inhalation of aerosols from
land application sites, consumption of groundwater affected by land-applied biosolids,
direct ingestion of biosolids-amended soils, consumption of water contaminated by
runoff from a land application site, and ingestion of plants impacted by land-applied
biosolids. Currently the model is able to quantify risks for six pathogens: Giardia lamblia,
Cryptosporidium, Salmonella, Shigella, enterovirus, and adenovirus, and examine the
exposure concentrations for four indicators: coliphage, E.coli, Enterococci, and fecal
coliforms.
vii
The application of the SMART Biosolids model to a specific site with typical
application found that the risks generated across pathways, in descending order, are from
ingestion of biosolids-amended soils, ingestion of contaminated surface water, ingestion
of contaminated vegetables, inhalation of aerosols from application sites, and
consumption of contaminated groundwater. A sensitivity analysis indicates that microbial
parameters, especially decay rates and dose-response parameters, are strongly correlated
to the risk estimates. For the groundwater pathway, the hydraulic parameters, including
hydraulic conductivity, saturated water content, residual water content, and dispersion,
need to match site-specific environmental conditions. This study compiles the most
current pathogen occurrence, fate, and decay data and develops a comprehensive
exposure model for biosolids-derived pathogens. The assessment tool has the capability
to archive the most up-to-date knowledge and to be updated as additional information
becomes available in the future.
KEY WORDS: Microbial risk assessment; Exposure model; Biosolids land
application; Subsurface fate and transport; Spreadsheet model
1
CHAPTER 1. INTRODUCTION
Biosolids, the treated sewage sludge resulting from wastewater treatment, has
been recycled as fertilizer to sustainably improve and maintain productive soils and to
stimulate plant growth for over forty years (U.S. EPA, 1994). There are approximately
3.4 million dry tons of biosolids applied to land annually in the United States (Pepper,
Brooks et al., 2010). While providing essential elements to improve soil structure,
biosolids may contain pathogens harmful to human health.
In 1993, the Environmental Protection Agency (U.S. EPA) established standards
for land-applied biosolids under 40 CFR Part 503 Rule, Standards for the Use or
Disposal of Sewage Sludge. The pathogen reduction standards specify pathogen limits for
two classes of biosolids on the basis of sludge treatment, and pathogen or indicator
organism content: Class A and Class B (U.S. EPA, 2000). Class A biosolids can be
applied on site without any pathogen-related restrictions; Class B biosolids contain trace
levels of pathogens and need further treatment before exposed to the natural environment.
The pollutant limits for pathogens in Class B biosolids were set based on a framework
stipulating that exposure to pathogens was to be reduced through treatment-based
standards or through land application guidelines, rather than on risk- or epidemiologically
based estimates. At the time when the regulation was adopted, microbial risk assessment
methodologies were not sufficiently developed to establish risk-based standards, and
sufficient exposure data were not available (U.S. EPA, 1989; U.S. EPA, 1992; U.S. EPA,
1995).
The U.S. EPA has used risk-based approaches for regulatory purposes for years.
Microbial risk assessment provides a scientifically based approach to characterize risks
even when risks are below the detection limits of epidemiological studies and exposure
2
measurements. With improved information on the pathogen and indicator content of
biosolids and advancements in the development of theoretically and empirically based
microbial transport models, it is now feasible to use the most current information on
microbial risk assessment to inform the management of biosolids land application
programs. Site-specific microbial risk assessment can be used to evaluate the suitability
of a specific site for biosolids application or to address risks if proper land application
procedures are not followed, thereby informing the development of response and
remediation plans. Risk assessments can also provide insight into the relative risks
associated with different pathogens and different exposure pathways.
1.1 Background
The characteristics and properties of biosolids vary depending on the quality and
origin of sludge, along with the type of treatment processes (Guzman, Jofre et al., 2007;
Viau and Peccia, 2009; Wong, Onan et al., 2010). In general, biosolids contain nitrogen,
phosphates, metals, organics, other elements, and microorganisms (Gerba, Pepper et al.,
2002; Overcash, Sims et al., 2005; Guzman, Jofre et al., 2007). During the 1970s and
’80s, the potential benefits and hazards of land application were studied in both the U.S.
and Europe (Pepper, Brooks et al., 2006). Biosolids have been shown to provide essential
elements to crops, and to increase soil organic content and improve soil structure.
Meanwhile, the soil environment helps stabilize potential pollutants from land-applied
biosolids (Mantovi, Baldoni et al., 2005; Harrison, Oakes et al., 2006). Pathogens
contained in biosolids include viruses, bacteria, and animal and human parasites
(protozoa and helminthes), which may cause various human diseases and illnesses (U.S.
EPA, 2000; Singh and Agrawal, 2008).
3
There are two classes of biosolids specified by the Part 503 Rule: Class A and
Class B. Class A biosolids are treated by one of several “Processes to Further Reduce
Pathogens” (PFRP), such as composting, pasteurization, drying or heat treatment, or
advanced alkaline treatment, which reduce pathogens to below detectable levels. Class B
biosolids are treated using a “Process to Significantly Reduce Pathogens” (PSRP), such
as aerobic digestion, anaerobic digestion, air drying, and lime stabilization, which reduce
but do not eliminate pathogens, and therefore other precautionary measures are required.
The Part 503 Rule specifies several site access and crop harvesting restrictions for Class
B biosolids, and several states impose even more stringent restrictions. The intent of the
Class B biosolids requirements is to ensure that biosolids can be safely land applied and
are unlikely to pose a threat to public health and the environment. Mesophilic anaerobic
digestion (MAD) is the most prevalent treatment process for Class B biosolids in the U.S.
with a mean reduction in pathogen or indicator cultivability of 1 log (Viau, Bibby et al.,
2011). It was found that indicator and pathogen levels within Class B biosolids have been
effectively reduced since the promulgation of the U.S. EPA Part 503 Rule in 1993
(Pepper, Brooks et al., 2010).
However, Part 503 Rule was not based on scientific methods to characterize
microbial risks and establish standards. Over the past decade, epidemiology-based health
investigations showing the association between health effects and biosolids have made
little progress. Meanwhile, there have been increased concerns about the adequacy of
the standards for protecting human health. In response to these concerns, the National
Research Council (NRC) was commissioned to independently review the scientific basis
of the regulation. They came to the conclusion that “additional scientific work is needed”,
4
and recommended the use of “improved risk-assessment methods to better establish
standards” (National Research Council, 2002).
1.2 Literature review
1.2.1 Microbial characterization of biosolids
Untreated wastewater may contain pathogens that can cause human infection and
illness (Table 1-1). Pathogen concentrations are significantly reduced in sludge which is
treated to meet the pathogen limits for Class A or Class B biosolids (Table 1-2). Biosolids
may have trace levels of different types of pathogens, such as bacteria, parasites, and
viruses, depending on the types of physicochemical- and biological-processes and the
extent of treatment used. The following section describes the occurrence of
microorganisms in Class B biosolids after treatment and before land application from a
number of studies.
The number of enteric viruses reported in raw and digested biosolids varies
widely from study to study due to the diverse methods for elution and quantification
(Sidhu and Toze, 2009). On the basis of available literature, norovirus numbers are 3 to 4
times higher than enteric virus and reovirus numbers in biosolids (Sidhu and Toze, 2009).
However, a different study found that norovirus concentrations were comparable to
enterovirus and polyomaviruses in Class B MAD biosolids (Wong, Onan et al., 2010).
Adenoviruses were reported to be 10 times higher than enteroviruses in wastewater
(Sidhu and Toze, 2009) and more prevalent than enteric viruses in Class B MAD
biosolids (Pepper, Brooks et al., 2010; Wong, Onan et al., 2010). No hepatitis A virus
was detected in Class B MAD biosolids (Wong, Onan et al., 2010).
5
Table 1-1 Major pathogens potentially present in municipal wastewater and manure (U.S.
EPA, 2000)
Bacteria Disease/Symptoms for organism
Salmonella spp. Salmonellosis (food poisoning), typhoid
Shigella spp. Bacillary dysentery
Yersinia spp. Acute gastroenteritis (diarrahea, abdominal
pain)
Vibrio cholerae Cholera
Campylobacter jejuni Gastroenteritis
Escherichia coli (enteropathogenic) Gastroenteritis
Viruses Disease/Symptoms for organism
Poliovirus Poliomyelitis
Coxackievirus Meningitis, pneumonia, hepatitis, fever,
etc.
Echovirus Meningitis, paralysis, encephalitis, fever,
etc.
Hepatitis A virus Infectious hepatitis
Rotavirus Acute gastroenteritis with severe diarrhea
Norwalk Agents Epidemic gastroenteritis with severe
diarrhea
Reovirus Respiratory infections, gastroenteritis
Protozoa Disease/Symptoms for organism
Cryptosporidium Gastroenteritis
Entamoeba histolytica Acute enteritis
Giardia lamblia Giardiasis (diarrhea & abdominal cramps)
Balantidium coli Diarrhea and dysentery
Toxoplasma gondii Toxoplasmosis
Helminth Worms Disease/Symptoms for organism
Ascaris lumbricoides Digestive disturbances, abdominal pain
Ascaris suum Coughing, chest pain
Trichuris trichiura Abdomen pain, diarrhea, anemia, weight
loss
Toxocara canis Fever, abdominal discomfort & muscle
aches
Taenia saginata Nervousness, insomnia, anorexia
Taenia solium Nervousness, insomnia, anorexia
Necator americanus Hookworm disease
Hymenolepis nana Taeniasis
Note: Not all pathogens are necessarily present in all biosolids and manures, all the time.
6
Table 1-2 The pathogen reduction standards (Part 503 Rule pathogen density limits) for
Class A and B biosolids (U.S. EPA, 2000)
Pathogen or indicator Standard density limits (dry wt)
Class A
Salmonella <3 MPN/4 g Total solids or
Fecal Coliforms <1000 MPN/g and
Enteric Viruses <1 PFU/4 g Total solids and
Viable Helminth Ova <1/4 g Total solids
Class B
Fecal Coliform <2,000,000 MPN/g Total solids (dry wt.
basis)
Indicator levels, including fecal coliform, E. coli, enterococci, and somatic phage,
were approximately 104 MPN or PFU in Class B MAD biosolids (Wong et al., 2010).
For bacterial pathogens, Salmonella numbers were reported high in raw sludge,
but are known to survive only in low numbers in biosolids (Sidhu and Toze, 2009; Pepper,
Brooks et al., 2010) and were positive in only two of six total Class B MAD samples with
concentrations below 1.0 MPN/4g (Wong, Onan et al., 2010). Escherichia coli O157:H7
in biosolids are not expected to be high, but they are known to survive in stored animal
manure for more than 11 weeks, and regrowth is possible under certain conditions. High
numbers of Campylobacter were reported in raw sewage sludge; Shigella was also
detected in Class B MAD biosolids (Pepper, Brooks et al., 2010).
For protozoan parasites, viable Cryptosporidium oocysts were present in most
treated sludges produced after mesophilic and thermophilic treatments (Guzman, Jofre et
al., 2007). One study suggests no statistically significant reduction in Cryptosporidium
oocysts or Giardia cysts during anaerobic sludge digestion, showing the persistence of
protozoa in sewage sludge (Chauret, Springthorpe et al., 1999); another study reports
7
higher reductions in Cryptosporidium numbers as compared to Giardia (Sidhu and Toze,
2009).
The available information for helminthes is exclusively on Ascaris, due to its
higher prevalence than other helminths, and because Ascaris eggs are resistant to
environmental conditions and remain infective for several years (Sidhu and Toze, 2009).
However, Ascaris ova were not detected in Class B biosolids across the United States by
Pepper et al. (2010) and were at very low levels in raw sludges (Guzman, Jofre et al.,
2007).
Joe (2011) has reviewed and compiled the published occurrence of pathogens and
indicators in Class B biosolids. The values are tabulated in Table A-1 in Appendix.
1.2.2 Fate and transport of microorganisms from biosolids
Microorganisms in biosolids have potential risk to human health by transport
through different exposure routes, including inhalation of aerosol from application site,
consumption of contaminated groundwater or surface runoff, ingestion of biosolids-
amended soils or contaminated vegetables. Groundwater is one of the primary drinking
water resources in the United State. Aquifers are traditionally viewed as effective natural
filters, which provide a buffer to protect underlying groundwater (Schijven, 2003;
Schijven et al., 2002). However, it has been reported that pathogenic contamination of
groundwater is responsible for many waterborne disease outbreaks (Macler et al., 2000;
Fout et al., 2003; Nasser et al., 1999; Pang, 2009).
There have been many models developed to predict the potential of contaminants
to reach the water table during infiltration (Corapcioglu et al., 1985; Ginn et al., 2002;
Pang, 2009; Sinton et al., 1997; Tufenkji, 2007; Waddill et al., 1998). Groundwater
8
transport models developed to simulate rainfall-induced infiltration describe the pathway
by which land-applied contaminants transfer through the subsurface to the underlying
source water (Curriero et al., 2001; Drayna et al., 2010; Risebro et al., 2007). Most
models are based on analytical and numerical solutions to the advection-dispersion
equation (Tufenkji, 2007). Fewer models take into account the effects of specific physical,
biological, and geochemical conditions. For example, in some cases, these models cannot
reflect the non-homogeneous and anisotropic nature of the subsurface system.
Microorganisms may travel much further through soil with root channels, rodent holes or
other macropores (Butler, 1954; Hagedorn, 1983; Li et al., 1996).
Numerous environmental factors have been identified to affect the transport and
survival of microorganisms in groundwater. Subsurface hydraulic parameters such as
local pore water velocity, dispersivity and filtering coefficients are very important in
determining microbial transport (Li et al., 1996). They vary widely with specific
subsurface properties, including physical, biological, and geochemical conditions (Tan,
1992) and may affect both the transport time and the decay rate for microorganisms.
Survival time of microorganisms in groundwater is another important factor
impacting their fate and transport, which significantly depends on temperature, especially
in areas with shallow aquifers. The survival times reported for pathogens and indicators
in the published literature have been reviewed and are compiled in Table A-2 of
Appendix. The persistence of pathogens in water is often reported using a pathogen
inactivation rate in units of log day-1
and following a log-normal distribution (Enriquez et
al., 1995; John and Rose, 2005; Cook and Bolster, 2007). Other papers reported the
percentage inactivation of pathogens at a given time (McFeters and Bissonnette, 1974;
9
Jackson et al., 1977; Griffiths, 1978; Feachem et al., 1983; Henis, 1987; Filip and Kaddu-
Mulindwa, 1988; Dubey, 1998; Medema et al., 1998; Koudela et al., 1999; Lyon and
Faulkner, 2001; Adams and Bates, 2003; Ramaiah et al., 2004; Erickson and Ortega,
2006; Azevedo and Almeida, 2008; Espinosa and Mazari-Hiriart, 2008; Ngazoa et al.,
2008; Kim and Jiang, 2010). Keswick et al. discuss the choice of indicator organisms for
viral pathogens based on their inactivation rates. The microorganisms were ranked in
order of decreasing decay rates as: coliphage f2, rotavirus SA-11, Escherichia coli,
echovirus-1, fecal streptococcus, poliovirus-1, and coxsackievirus B3 (Keswick et al.,
1982). E-coli (coliforms and fecal coliforms) had the fastest die-off rate making it a less
sensitive indicator than fecal streptococcus. Enterococcus, a fecal streptococcus that is
typically more human-specific than the larger fecal streptococcus group, appears to be a
better indicator of potential disease hazards in groundwater and other waters than
coliform and fecal coliform organisms. EPA recommends Enterococci as the best
indicator of health risk in salt water used for recreation and as a useful indicator in fresh
water as well (EPA, 2011).
There are few published research reports on the fate and transport of pathogens
from biosolids in the environment. The potential release of viruses and indicators from
biosolids to the aqueous phase was investigated recently (Table A-3 in Appendix)
(Chetochine et al., 2006; Xagoraraki et al., 2010). Results indicate that less than 8% of
the pathogens or indicators were released from the biosolids-soil matrix. Bitton et al. did
not find any poliovirus or echovirus in soil leachates collected after natural rainfall
(Bitton et al., 1984). Horswell et al. observed a significant difference in the release
fraction from sewage sludge leachate between Salmonella (30%) and adenovirus (0.08%).
10
1.2.3 Dose-response relationship
The virulence of the disease from the infectious agent is another essential input to
microbial risk assessment models. Researchers have studied the relationship between the
ingested dose and the resulting risk to human health, which is described by dose-response
parameters (Regli et al., 1991; Teunis, 1996; Haas et al., 1999; Soller et al., 2004; Lathem
et al., 2005; Armstrong and Haas, 2007; Asano, 2007; Smith et al., 2008; Bouwknegt et
al., 2009; Chacin-Bonilla, 2010; Mara and Sleigh, 2010). Literature reported values are
compiled in Table A-4 in Appendix.
1.2.4 Microbial risk assessment for land-applied biosolids
To date, several quantitative microbial risk assessment (QMRA) studies have
reported the risks from biosolids to the residential public, focusing on the exposure
scenarios of accidental direct ingestion, aerosol inhalation, groundwater direct ingestion,
and contaminated food ingestion (Dowd et al., 2000; Brooks et al., 2004; Westrell, 2004;
Brooks et al., 2005; Brooks et al., 2005; Gale, 2005; Eisenberg et al., 2008). Figure 1-1
shows annual infectious risks from different exposure pathways for different pathogens
(Viau et al., 2011). It is found that the accidental direct ingestion produced the highest
annual risk, inhalation produced the next highest risk, and that risks from groundwater
and direct ingestion of contaminated food were low.
11
Figure 1-1 Literature values for annual infective risks (in log10 scale) from land-applied
biosolids through different exposure routes (Viau, Bibby et al. 2011)
Of all the published QMRA studies, only Eisenberg et al. (2006) developed and
demonstrated a microbial risk assessment framework for biosolids-associated pathogens
through consumption of groundwater. However, there are several challenges associated
with implementing this framework in a site-specific assessment. They are as follows:
(1) Eisenberg et al. identified exposure scenarios of groundwater and surface
water; however, the exposure models did not evaluate wet weather events. The risk of
exposure to biosolids-associated pathogens through ground water and surface water
would be enhanced under wet-weather conditions because pathogens in biosolids may
infiltrate down to the water table with stormwater infiltration, or may be carried to nearby
ponds and streams with stormwater runoff.
12
(2) For the groundwater exposure model, Eisenberg et al. considered three distinct
types of media: non-porous media (karst/bedrock), unsaturated soil and saturated soil.
They assumed several fixed thicknesses for each homogenous soil layer and an overall
attenuation for the heterogeneous scenarios was obtained by summing the predicted log
removals in each layer. The mechanisms governing microorganism fate and transport
were simplified based on these assumptions without further evaluation.
(3) Eisenberg et al. only considered viral pathogens, including coxsackievirus,
echovirus, Hepatitis A virus, poliovirus, and rotavirus, and did not provide any guidance
on bacteria or parasites, which also occur in biosolids and pose risks to human health.
(4) Eisenberg et al. identified five major exposure pathways: (a) inhalation of
aerosols from land application sites, (b) consumption of groundwater impacted by land-
applied biosolids, (c) direct ingestion of biosolids-amended soils, (d) ingestion of plants
impacted by land–applied biosolids, and (e) consumption of water contaminated by
runoff from a land application site. However, they only conducted risk assessment
methodology for the first three exposure pathways and there was no comparison of risks
across the different pathways.
1.2.5 Development of spreadsheet tools
There are a number of applications of spreadsheets in microbial risk assessment,
most of which are in the area of food safety (Vose, 1998; Ross and Sumner, 2002;
Vandeven et al., 2002; Hutter and Kihm, 2010). These tools provide an estimate of the
most probable outcome, but most do not provide information about the level of
confidence or the probable range of infections and illnesses for different scenarios (Ross
and Sumner, 2002). Some simulation models provide an uncertainty analysis by using a
13
spreadsheet combined with an add-in computer program, such as @Risk or Crystal Ball
(Vose, 1998; Lindqvist and Westoo, 2000; Oscar, 2002; Oscar, 2004; Carrasco and
Chang, 2005).
Several studies use spreadsheets to develop part of the quantitative microbial risk
assessment, such as modeling initial concentrations (Carrasco and Chang, 2005), dose-
response relationship (Teunis et al., 2010), pathogen transmission dynamics (Nauta,
2005), and risk ranking (Sumner and Ross, 2002). While there are several well-developed
spreadsheet-based environmental fate and transport models (Park and San Juan, 2000;
Rucker, 2007; Dixon et al., 2008; Knightes, 2008), to date, there are no available
comprehensive spreadsheet models that link quantitative microbial risk assessment with
microbial fate and transport modeling.
1.3 Research objectives
The goal of this study is to develop and apply a site-specific microbial risk
assessment tool for biosolids in order to provide a technical basis for assessing human
health risks that result from consumption of ground water contaminated by biosolids-
associated pathogens, and to compare this risk to other exposure pathways. This research
will lead to four journal papers, focusing on 1) incorporation of wet-weather events into
the transport model (Chapter 2 of this thesis, published by Journal of Hydrologic
Engineering in March 2012), 2) development and field validation of the subsurface fate
and transport model (Chapter 3 of this thesis), 3) encoding of the SMART Biosolids
model in a spreadsheet environment (Chapter 4 of this thesis, submitted to Environmental
Modeling & Software), and 4) application and interpretation of the SMART Biosolids
model (Chatper 5 of this thesis).
14
1.3.1 Objective 1: Model development - addressing wet-weather events
The occurrence of rainfall events has been significantly linked to waterborne-
disease outbreaks, indicating that wet-weather events may have significant impacts on
microbial risk (Rose et al., 2000). An important step in developing a QMRA for exposure
to biosolids-associated pathogens through ground water is to characterize the risk of
storm-induced infiltration and runoff, the hydrologic processes most likely to introduce
soil amendment-associated pathogens to source water. Objective 1 is to develop a method
to determine the probabilities of different amounts of infiltration and runoff based on
commonly available precipitation intensity duration and frequency curves. This approach
is then applied to a case study in order to examine infiltration depth and runoff volume
for wet-weather events associated with a range of occurrence frequencies. This research
provides a sound method to determine the storm events with the greatest potential for
mobilization of pathogens in the environment and the greatest risk to human health. This
analysis is presented in Chapter 2, and journal paper on this topic has been published by
Journal of Hydrologic Engineering (Teng et al., 2012a).
1.3.2 Objective 2: Model development - subsurface fate and transport of pathogens
The exposure model is the most essential and complicated part of QMRA. Both
model structure and parameter selection are crucial for the reliability of the results. The
subsurface fate and transport model must take both hydraulic and microbial parameters
into account, including flow rate, dispersivity, retardation, and microbial decay rate.
Objective 2 is to develop and validate the microbial subsurface fate and transport model,
based on field monitoring studies. Indicator and pathogen relationships were identified
and discussed. This analysis presented in Chapter 3, and journal paper on this topic is in
15
preparation (Teng et al., 2012b).
1.3.3 Objective 3: Model integration - spreadsheet environment
Due to the fact that risk assessments usually consist of multiple linked modules
(e.g. exposure assessment, dose-response, risk characterization), each with their own set
of assumptions, inputs, and computations, risk assessment models tend to be dense and
poorly documented, making it difficult for others to reproduce a risk assessment (Vose,
1998; Kopylev et al., 2007; Arunraj and Maiti, 2009; Choun and Elnashai, 2010; Donald
et al., 2011). The development and use of a simple tool for microbial risk assessment
allows the risk assessment to be understandable and reproducible. Embedded macros,
which are spreadsheet add-ins, can perform repeated iterative computations in a much
more efficient way while maintaining a familiar user interface. Objective 3 is to integrate
available knowledge from diverse sources to an environmental dispersion, exposure, and
risk model, named the Spreadsheet Microbial Assessment of Risk: Tool for Biosolids
(“SMART Biosolids”). An application for a site-specific scenario is also provided to
demonstrate the model’s benefits and limitations. This analysis is presented in Chapter 4,
and journal paper on this topic is under second round review by Environmental Modeling
& Software (Teng, et al., 2012c).
1.3.4 Objective 4: Model application – uncertainty analysis and sensitivity analysis
The microbial risk assessment tool for land-applied biosolids must take several
crucial elements into account, including occurrence values of pathogens in biosolids, the
potential routes of infection, the probability of human exposure to the source of the
pathogen, as well as the amount that humans would ingest, and the virulence of the
infectious agent (U.S. EPA, 2000). Uncertainties related to each of these factors will
16
affect the confidence on the final prediction of risk. Objective 4 is to use uncertainty
analysis to compare the cumulative risks over time across organisms and exposure
pathways, and to use sensitivity analysis to identify the most important sources of
uncertainty for the predictions of risks. The results of this analysis are presented in
Chapter 5.
17
CHAPTER 2. MODEL DEVELOPMENT - ADDRESSING WET-WEATHER
EVENTS
2.1 Introduction
The issue of how to address wet-weather events in quantitative microbial risk
assessments of pathogens in the environment is an important challenge that has not been
well addressed in the literature. For groundwater, the saturated wetting front produced by
rainfall infiltration dramatically advances the transport of land-applied pathogens through
the unsaturated zone, often forming a direct connection to the water table. Rose et al.
(2000) confirmed a statistically significant relationship between precipitation events and
waterborne-disease outbreaks originating from groundwater sources, suggesting that
infiltrating wetting fronts are responsible for mobilizing and transporting pathogenic
organisms (Rose et al., 2000). To date, most studies of pathogen transport in unsaturated
soil have been conducted under steady-state flow conditions (Lance and Gerba, 1984;
Powelson and Gerba, 1994; Chu et al., 2001; Torkzaban et al., 2006; Van Cuyk and
Siegrist, 2007; Kenst et al., 2008), which rarely occur in nature. Under more realistic
transient conditions, a zone of near saturation forms behind the wetting front and
transport takes place through a near-saturated transmission zone (Kenst et al., 2008).
Kenst et al. (2008) have recently shown that virus transport during infiltration of a
wetting front is similar to that during steady-state saturated flow.
Many scenarios used for risk assessment of groundwater contamination fail to
account for pathogen transport during rainfall events in which rainwater advancing via a
saturated infiltrating wetting front may significantly reduce or even eliminate the
unsaturated buffer zone. For surface water, precipitation may cause contaminated runoff
from agricultural fields to enter drinking-water sources and recreational waters. In both
18
cases, accurate predictions of expected infiltration depth or runoff volume resulting from
a given rainfall event and their probabilities of occurrence, are imperative for assessing
risk of pathogen transport via groundwater or surface water, respectively.
Predicted risk on the basis of exposure to water contaminated by fertilizer-
associated pathogens through groundwater or surface water is conditional on the
probabilities of infiltration or runoff generated from wet-weather events. A rain event of
varying intensity and duration applied over a range of soil types will produce different
saturated infiltration depths and runoff volumes. To measure the overall risk, the
probabilities of different amounts of infiltration and runoff need to be determined.
However, historical statistics on infiltration and runoff quantities are not directly
available. Instead, a large database of precipitation intensity duration and frequency
curves for the United States is available (Frederick et al., 1977; Fernández et al., 1999;
Trefry et al., 2005; Gerold and Watkins, 2005; Guo, 2006; Singh and Zhang, 2007). In
this chapter we present a method to determine the probabilities of different amounts of
infiltration and runoff on the basis of these commonly available intensity-duration-
frequency curves. We then apply this approach to a case study to examine the effects of
soil texture and precipitation duration and intensity on infiltration depth and runoff
volume for wet-weather events associated with a range of occurrence frequencies.
2.2 Methods
The proposed approach involves first predicting the infiltration and runoff
generated in different soil types from rainfalls of various durations and intensities
expected for a given return period. The combination of rainfall intensity and duration that
produces the maximum infiltration for a given return period is then used to define the
19
critical infiltration value for that return period (i.e., the maximum of the infiltration
values produced by any of the rainfall events expected within that return period).
Similarly, for surface runoff, the combination of rainfall intensity and duration for a
given return period that produces the maximum runoff represents the critical runoff
volume associated with that return period. The same rainfall event would probably not be
responsible for producing both the maximum infiltration and maximum runoff for a given
return period. Maximum runoff would typically be produced by a high intensity event,
which would likely be of short duration. Maximum infiltration would likely be produced
by a longer duration, lower intensity event, which favors prolonged infiltration over
runoff. The following sections detail the application of this method to a case study in
Ingham County, Michigan.
2.2.1 Frequency analysis of historical precipitation records
A series of generalized precipitation-frequency maps were analyzed to generate
rainfall depth data for different rainfall durations (5, 10, 15, and 30 min, and 1, 2, 3, 6,
12, 18, and 24 h) over a range of return periods (2, 5, 10, 25, 50, and 100 years) for
Ingham County, Michigan. Each return period is associated with several precipitation
magnitudes and durations. U.S. Weather Bureau Technical Paper No. 40 presents
precipitation-frequency values for durations from 30 min to 24 h on the basis of data
from 200 first-order Weather Bureau stations which maintained complete recording-
gauge records (Hershfield, 1961). The National Oceanic and Atmospheric Administration
Technical Memorandum (NWS HYDRO-35) provides 5- to 60-min precipitation
frequencies for the 37 eastern and central states (Frederick et al., 1977). Precipitation data
was initially reported for return periods of 2 and 100 years and durations of 5, 15, and 60
20
min, which were used to calculate the data gaps of intermediate durations (Dx) (Equations
2-1 and 2-2) and intermediate return periods (Ty) (Equations 2-3 to 2-6) using the
weighted average approach of Frederick et al. (1977). Trefry et al. (2005) reported
precipitation data for durations from 1-24 h (Trefry et al. 2005). All the reported and
calculated precipitation data are tabulated in Table 2-1.
Precipitation depth for 10- and 30-min durations (Frederick et al., 1977)
10 15 5D 0.59D 0.41D (2-1)
30 60 15D 0.49D 0.51D (2-2)
where Dx = precipitation depth for the x-min duration storm.
Precipitation depth for intermediate return periods (Frederick et al., 1977)
5 100 20.278 0.674T T T (2-3)
10 100 20.449 0.496T T T (2-4)
25 100 20.669 0.293T T T (2-5)
50 100 20.835 0.146T T T (2-6)
where Ty represents precipitation depth for the y-year return period storm.
These precipitation events with known frequencies can be combined with rainfall
intensities to develop generalized intensity-duration-frequency relationships (Figure 2-1).
The probability, p, of a wet-weather event occurring in any given day can be calculated
from its return period, n years, using:
1
365p
n
(2-7)
21
Table 2-1 Calculated precipitation values (cm) for Ingham County (Michigan, U.S.A.)
Duration (h) Return period (years)
2 5 10 25 50 100
0.0833
(or 5 min.)
1.05
(1.02,1.08) a
1.25
(1.21,1.29) b
1.40
(1.36,1.45) b
1.62
(1.57,1.68) b
1.79
(1.73,1.85) b
1.97
(1.91,2.02) a
0.1667
(or 10 min.)
1.55
(1.47,1.64) b
1.92
(1.83,2.00) b
2.18
(2.09,2.27) b
2.54
(2.46,2.64) b
2.84
(2.74,2.95) b
3.12
(3.02,3.23) b
0.25
(or 15 min.)
1.91
(1.78,2.03) a
2.38
(2.26,2.50) b
2.72
(2.59,2.84) b
3.20
(3.07,3.30) b
3.56
(3.45,3.68) b
3.94
(3.81,4.06) a
0.5
(or 30 min)
2.59
(2.40,2.77) b
3.30
(3.10, 3.48) b
3.78
(3.60, 3.99) b
4.50
(4.29,4.70) b
5.05
(4.83,5.26) b
5.59
(5.36,5.82) b
1 3.30
(3.05,3.56) a
4.27
(3.99,4.55) b
4.93
(4.65,5.18) b
5.84
(5.56,6.15) b
6.58
(6.27,6.88) b
7.31
(6.985,7.62) a
2 2.92 c 3.73
c 4.42
c 4.98
c 5.69
c 6.40
c
3 2.84 c 3.63
c 4.29
c 4.88
c 5.56
c 6.30
c
6 2.77 c 3.53
c 4.17
c 4.72
c 5.41
c 6.15
c
12 2.77 c 3.48
c 4.09
c 4.62
c 5.28
c 5.97
c
18 2.77 c 3.48
c 4.06
c 4.60
c 5.23
c 5.92
c
24 3.12 c 3.89
c 4.55
c 5.11
c 5.82
c 6.60
c
a Values were obtained from Frederick, Myers et al. (1977).
b Values indicate average precipitation values with lower and upper bound precipitation
values shown in parentheses, as calculated using the equations given by Frederick, Myers
et al. (1977). c Values were obtained from Trefry et al. (2005).
22
Figure 2-1 Intensity-duration-frequency relationships for Ingham County (Michigan,
U.S.A.) at different return periods (Source: Frederick, Myers et al. 1977, Trefry et al.,
2005)
*Values on both axes are plotted on logarithmic-scales.
2.2.2 Infiltration and runoff modeling
A fully explicit infiltration model (Zhang et al., 2009) was developed to predict
infiltration depth and runoff volume associated with each defined precipitation event. The
joint Green-Ampt infiltration model was created by combining a constant flux Green-
Ampt model (Swartzendruber, 1974) with an explicit approximation to the Green-Ampt
model (Green and Ampt, 1911, Salvucci and Entekhabi, 1994). Taking into account
evidence that infiltration patterns change when surface soil is saturated (Chiu et al.,
2009), the constant flux Green-Ampt model is used up until the time of saturation, and
the explicit Green-Ampt model is used for infiltration estimates after the time of
saturation. A validation of this technique is provided in Zhang et al. (2009). Surface
23
runoff is then calculated as the excess of rainfall over infiltration. This model is
appropriate for conditions when its two underlying models are appropriate: a
homogeneous soil profile and no surface ponding.
The joint Green-Ampt model has the following form:
When r<Ks, and when both r>Ks and t<to
q r (2-8)
I rt (2-9)
When r>Ks and t>to
1/ 2 1/ 22 2 2 1 2( )
2 3 6 3s
q K
(2-10)
2 22 2 2 2
1 1 ln ln ln ln / 23 3 3 3 2
sI K t t t t t t t
(2-11)
0
Z
s
I
(2-12)
With
0( )( )
s f s
s
h h
K
(2-13)
t
t
(2-14)
0
0
( )
( )
s f s
s
K ht
r r K
(2-15)
where q = surface infiltration rate (cm=h); I = cumulative infiltration (cm); Z =
wetting front depth (cm); r = constant water application rate at the surface (cm=h); t =
24
time (h); Ks = saturated hydraulic conductivity (cm=h); θs = saturated volumetric water
content (cm3 =cm3 ); θo = initial volumetric water content (cm3=cm3); hf = capillary
pressure head (< 0) at the wetting front (cm); hs = ponding depth or capillary pressure
head at the surface (cm); and to = time when surface saturation occurs (h). The capillary
pressure is determined by
1f e
h h
(2-16)
with
2 3 (2-17)
1
2e b
h h (2-18)
where he = air exit head (cm); hb = air entry head (cm); λ = Brooks-Corey water
retention constant; and η = Brooks-Corey conductivity constant. When no measurement
is available, θr, the residential volumetric water constant, can be used in place of θo.
Typical values for θs, θr, hb, λ, and Ks for a range of soil types are provided by U.S. EPA
Report 600/R-97/128b (U.S. EPA, 1998). To develop conservative (upper bound)
estimates of infiltration and runoff, different soil hydraulic property parameters were
selected to model infiltration and runoff. In this paper, soil properties from Carsel and
Parrish (1988) were chosen for infiltration and data from Pajian (1987) was used for
runoff, as these parameters maximized the amount of infiltration and runoff, respectively.
2.2.3 Determination of critical rainfall
Results from the infiltration and runoff model can be used to determine the
critical infiltration and runoff volume for each return period. The critical infiltration is
defined by the rainfall intensity and duration that produces the maximum infiltration for
25
groundwater transport; the critical runoff is defined by the rainfall intensity and duration
that produces the maximum runoff volume.
2.3 Results
2.3.1 Comparison of predicted infiltration depth with experimental data
To assess the accuracy of this approach, comparisons were made between the
predicted infiltration depths and data from previously published laboratory experiments
(Fohrer and Berkenhagen, 1999). The experiments were performed with a capillary
rainfall simulator applying rainfall with the intensity of 3 cm/h continuously for 2 h. The
changes in water content during rainfall were observed using time-domain reflectometer
(TDR) field probes at three depths (3, 16, and 21 cm). The initial water content was
estimated to be 12% at depth of 10 cm, 20% at 20 cm, and 28% at 30 cm (volume basis).
The infiltration front reached a depth of 10 cm after 40 min of rainfall, and the water
content attained a maximum value around 50 min. The water front reached a depth of 20
cm after a little over 50 min and the maximum water content value was attained after 80
min. The wetting front arrived at 30 cm after a little over 90 min and the water content
kept increasing until 120 min. The maximum value of water content attained was roughly
36.4%.
Model predictions were compared with observations for the three different depths,
10, 20, and 30 cm. When available, site-specific parameters for the lab experiments were
used for the model inputs (Table 2-2). Fohrer and Berkenhagen (1999) did not report a
hydraulic conductivity but did report that silt-clay-loam soil was used. A lognormal
distribution was used to represent the range of values for hydraulic conductivity of silt-
clay-loam soil reported by Carsel and Parrish (1988). The initial water content may
26
change during a continuous rainfall event. A uniform distribution was used to represent a
plausible range of variation for initial water content values, with the lower bound set at
the residual water content of the dry silt-clay-loam soil (Carsel and Parrish, 1988) and the
upper bound set at the maximum value of water content attained in the laboratory
experiments. Table 2-3 shows the results of a Monte Carlo analysis (1000 trials)
comparing predicted and observed infiltration depth.
Table 2-2 Site-specific parameters for lab experiments study
Parameters Unit Distribution type Values
Soil texture Silt clay loam
Hydraulic conductivity cm/h Log-normal 0.07 (0.07, 0.19)a
Initial water content cm3/
cm3 Uniform 0.23 (0.089, 0.364)
b
Saturated water content cm3/
cm3 Constant 0.364
a (mean, standard deviation)
b (min, max)
27
Table 2-3 Comparison between the predicted and observed infiltration depth
Duration (min) Predicted (cm) Observed (cm)
Scenario 1 50 11.97 (6.05, 26.81) a 10
Scenario 2 80 17.13 (7.71, 35.21) a 20
Scenario 3 90 23.53 (9.53, 50.86) a 30
a (5%, 95% confidence intervals)
The observed infiltration depths for all three durations fell into the 90% predicted
ranges from the Monte Carlo analysis. The mean infiltration depth predicted for Scenario
1 was roughly 2 cm higher than the observed depth, which may result from the
uncertainty within the estimates of the hydraulic parameter values. The soil texture
described in the experimental paper was a typical loess-derived soil with 63.3% silt, 29.5%
clay, and 7.2% sand, whereas the default hydraulic parameter values in our prediction
model are for silt-clay-loam soil with 56% silt, 34% clay, and 10% sand. Because the
experimental soil had more silt, more clay, and less sand than the soil for which Carsel
and Parrish (1988) developed their conductivity estimates, the model may somewhat
overestimate the infiltrated water. However, predictions of mean infiltration depth for
Scenarios 2 and 3 were underestimated by 3–7 cm. This may be attributable to the input
values for initial water content. In reality, the water content is higher for a deeper location
and increases with time, whereas the water content was assumed to be constant in the
modeling, which would underestimate the soil infiltration capacity.
2.3.2 Infiltration and runoff predictions
Calculated values for total precipitation associated with 2-, 5-, 10-, 25-, 50-, and
100-year storms for Ingham County, Michigan, are shown in Table 2-1. For each return
period, storm precipitation values were calculated for durations of 5, 10, 15, and 30 mins,
as well as 1, 2, 3, 6, 12, 18, and 24 h. These storm values were used as input to the
28
infiltration and runoff model to predict infiltration depth and runoff volume for each
storm.
Figure 2-2 shows values of wetting front depth and surface runoff produced by
loam soil over the range of rainfall durations and intensities representing a 100-year
return period rainfall. For the 100-year storm, there are several non-runoff-producing
rainfalls, which consequently produce maximum infiltration. High infiltration is produced
by a lower intensity (and therefore longer duration) storm. When rainfall intensity is
lower than the soil infiltration capacity, water infiltrates at the same rate as the rainfall
intensity; however, when rainfall intensity is greater than the infiltration capacity, excess
water becomes runoff once the soil is saturated. Wetting front depth increases initially as
longer duration storms generate more total infiltration (even if they are less intense), but
then tends to plateau as the longer duration events correspond with lower intensity
rainfalls. A peaked relationship is observed between runoff and rainfall intensity or
duration. High intensity storms tend to have short durations and produce little total runoff.
Low intensity storms (which tend to be longer in duration for the same return period) do
not exceed the infiltration capacity of the soil. Thus the greatest runoff is produced by
storms of intermediate intensity and duration.
29
Figure 2-2 Infiltration (i.e., wetting front depth) and runoff (i.e., surface runoff) produced
by loam soil for a 100-year return period rainfall in relation to rainfall duration and
intensity
Figure 2-3 shows the wetting front depth and surface runoff produced by different
soil texture classes under a 100-year return period rainfall. As expected, coarse sand with
the highest infiltration capacity produces the greatest infiltration and the lowest runoff.
Maximum runoff is produced by clay. Figure 2-4 shows the wetting front depth and
surface runoff produced by loam soil under all the rainfall events with different return
periods. Both wetting front depth and runoff depth increase with more severe, less
30
frequent storm events.
31
Figure 2-3 Infiltration (i.e., wetting front depth) and runoff (i.e., surface runoff) produced
by different soil texture classes for a 100-year return period rainfall in relation to rainfall
duration and intensity
32
Figure 2-4 Infiltration (i.e., wetting front depth) and runoff (i.e., surface runoff depth)
produced by loam soil in relation to rainfall duration and intensity
2.3.3 Determination of critical rainfall
Predictions of infiltration and runoff produced in loam soil for rainfalls
representative of the 100-year storm are shown in Table 2-4. The maximum wetting front
depth is produced from the 2-h and 3.2-cm/h rainfall event and the maximum runoff is
produced from the 1-h and 7.3-cm/h rainfall event. Maximum infiltration does not
immediately occur once the rainfall intensity exceeds the loam’s infiltration capacity of
1.32 cm/h, because water may infiltrate at a rate higher than the saturated infiltration
33
capacity before the soil is saturated. Maximum infiltration is observed for a 2-h storm
with an intensity of 3.2 cm/h, the most intense storm that does not produce runoff. High-
intensity rainfalls saturate soil more quickly than low-intensity rainfalls. For the 7.3-cm/h
rainfall, the soil saturates in less than 1 h, thereby producing runoff. Whereas runoff is
also produced during higher intensity rainfalls, rainfall duration decreases with increasing
intensity for storms of similar return period so maximum runoff occurs during the 7.3-
cm/h rainfall.
Table 2-4 Predictions of infiltration and runoff values produced in loam soil for 100-year
return period rainfall
Precipitation values Infiltration and runoff predictions
Precipitation
depth (cm)
Duration
(h)
Intensity
(cm/h)
Wetting front depth
(cm)
(Groundwater
transport scenario)
Runoff (cm)
(Surface water
transport scenario)
1.97 0.08 23.62 3.66 0.80
3.13 0.17 18.78 5.31 1.43
3.94 0.25 15.75 6.62 1.81
5.59 0.50 11.17 9.74 2.45
7.30 1.00 7.30 14.55 2.59
a
6.40 2.00 3.20 22.14
a
0.00
6.30 3.00 2.10 17.90 0.00
6.15 6.00 1.02 17.46 0.00
5.97 12.00 0.50 16.96 0.00
5.92 18.00 0.33 16.81 0.00
6.60 24.00 0.28 18.76 0.00 a Bold values indicate maximum values.
Critical rainfall values and characteristics (magnitude, duration, intensity)
34
producing maximum infiltration and runoff were defined over a range of return periods,
as shown in Table 2-5. Also included in Table 2-5 are daily probabilities of occurrence of
the critical infiltration- and runoff-producing rainfall events. Corresponding infiltration
and runoff depths for each return period are plotted in Figure 2-4. Results presented in
this chapter are for one case study only, but the approach may be applied at other
locations using the appropriate soil texture class and intensity-duration-frequency curves.
Table 2-5 Critical rainfall values for runoff and infiltration in loam soil for surface water
and groundwater transport scenarios
Return
period
(year)
Probability of
occurrence in
one day
Surface water transport Groundwater transport
Critical
rainfall
duration
(h)
Critical
rainfall
intensity
(cm/h)
Runoff
produced
by critical
rainfalls
(cm)
Critical
rainfall
duration
(h)
Critical
rainfall
intensity
(cm/h)
Wetting
front
depth
produced
by
critical
rainfalls
(cm)
2 1.37×10-3
0.25 7.62 0 0.5 5.18 9.74
5 5.48×10-4
0.25 9.51 0.25 1 4.26 14.55
10 2.74×10-4
1 4.92 0.65 1 4.92 14.55
25 1.10×10-4
0.5 8.99 1.35 1 5.36 14.55
50 5.48×10-5
0.5 10.08 1.90 2 2.84 16.16
100 2.74×10-5
1 7.3 2.59 2 3.20 22.14
35
2.4 Discussion
This method predicts the maximum infiltration and runoff depths for a known-
frequency wet-weather event. Because the risk of human exposure to pathogens
originating from land-applied soil amendments increases dramatically with infiltration to
the water table and surface water runoff, this information is critical to the development of
comprehensive microbial risk assessments of land-application practices. By coupling the
maximum infiltration and runoff depths resulting from the critical rainfalls with their
probabilities of occurrence (Table 2-5), it is possible to estimate the probability of
creating a saturated connection to the water table or producing surface water runoff of a
given amount. The daily probabilities of rainfall events capable of producing
contaminated runoff is 5.48 × 10-4
, and the daily probabilities of contaminating
groundwater with 0.2 m water table depth is 2.74 × 10-5
. This information about the
probability of different infiltration depths and runoff amounts may be coupled with
models of environmental transport, exposure, and risk scenarios to determine pathogen
concentrations and human health risks given the occurrence of particular rainfall events
(Eisenberg et al., 2006; Zhang et al., 2009; Teng et al., 2010; Kumar et al., 2010).
These results indicate that wet-weather modeling cannot be based simply on
selecting a single storm to characterize a given return period. A 100-year storm may
produce no runoff or up to 2.59 cm of runoff and create a saturated wetting front ranging
from 3.66–22.14 cm, depending on the rainfall duration and intensity. To avoid
underestimating human health risks associated with pathogen transport from land-applied
soil amendments, it is important to evaluate a range of storms for a given return period to
identify the highest potential for environmental transport.
36
CHAPTER 3. MODEL DEVELOPMENT - SUBSURFACE FATE AND
TRANSPORT OF PATHOGENS
3.1 Introduction
Waterborne microorganisms of public-health concern can enter aquifers via
several sources, including sinkholes, septic systems, rain infiltrating through landfills, or
surface water through wells. Land application of biosolids is an additional potential
source for microbial contamination in groundwater systems (Corapcioglu, 1985; Pang,
2009). Under certain conditions, microorganisms originating from the biosolids may
move through the soil and enter the underlying groundwater supply (Corapcioglu, 1984;
Romero, 1970; Butler, 1954; Hagedorn, 1983).
In order to improve understanding of the fate and transport of microbes in the
subsurface, there have been many models developed to predict the potential of
contaminants to reach the water table during infiltration (Corapcioglu et al., 1985; Ginn
et al., 2002; Pang, 2009; Sinton et al., 1997; Tufenkji, 2007; Waddill et al., 1998).
Groundwater transport models developed to simulate rainfall-induced infiltration describe
the pathway by which land-applied contaminants transfer through the subsurface to the
underlying source water (Curriero et al., 2001; Drayna et al., 2010; Risebro et al., 2007).
Most models are based on analytical and numerical solutions to the advection-dispersion
equation (Tufenkji, 2007). Numerous environmental factors affect the transport and
survival of microorganisms in groundwater. Subsurface hydraulic parameters such as
local pore water velocity, dispersivity and filtering coefficients are very important in
determining microbial transport (Li et al., 1996). They vary widely with specific
subsurface properties, including physical, biological, and geochemical conditions (Tan,
1992) and may affect both the transport time and the decay rate for microorganisms.
37
Survival time of microorganisms in groundwater is another important factor impacting
their fate and transport, which also varies with temperature, especially in areas with
shallow aquifers.
Fewer models take into account the effects of specific physical, biological, and
geochemical conditions. For example, in some cases, these models cannot reflect the non-
homogeneous and anisotropic nature of the subsurface system. Microorganisms may
travel much further through soil with root channels, rodent holes or other macropores
(Butler, 1954; Hagedorn, 1983; Li et al., 1996). There is also no fate and transport model
calibrated specifically to data for the transport of biosolids-associated indicators and
pathogens in the subsurface. Indicators and pathogens from biosolids may differ in
several respects from typical organisms. For example, these organisms might be less
prone to sorption, as they are selected to be organisms that have not remained sorbed to
the biosolids, and may travel farther in soil. In addition, the leaching of microorganisms
from biosolids needs to be estimated, which is not a parameter required in other
subsurface fate and transport modeling applications.
In this study, a subsurface fate and transport model was developed and calibrated
to data on transport of indicators from land-applied biosolids. Different model
components, including an infiltration model, a saturated transport model and an
unsaturated transport model, were validated individually using published data from
laboratory-based column studies (Galada et al., 2012a). In this chapter, effluent samples
from field lysimeter studies were used to validate pathogen concentration predictions
from the fate and transport model, and to estimate field parameters that best characterize
the subsurface transport.
38
3.2 Methods
3.2.1 Subsurface fate and transport model
A subsurface fate and transport model is linked to a one-dimensional infiltration
model with a groundwater transport model (Galada et al., 2012a) to calculate subsurface
concentrations of biosolids-derived pathogens following biosolids application and a
subsequent wet-weather event. The infiltration model, which is based on a Joint Green-
Ampt model (Galada et al., 2012), predicts the wetting front depth due to rainfall events
(Teng, Kumar et al., 2012a). The thickness of the unsaturated soil below the wetting front
and the saturated soil above the wetting front provides boundary information for the
microbial transport and fate model. By comparing the wetting front depth to water table
depth, two transport scenarios are considered (Figure 3-1). A “saturating event” occurs
when the infiltrating wetting front from the rainfall event saturates through to the ground
water table, creating a fully saturated connection. In this case microorganisms vertically
transport through the saturated soil and join the horizontal saturated groundwater flow
without any attenuation through non-saturated soil. When the rainfall is not large enough
to saturate all the soil above the water table, a “non-saturating event” occurs in which
there is vertical transport both through saturated soil and through an unsaturated layer. A
“saturating event” results in greater transport of pathogens. Based on the boundary
information calculated from infiltration models from Chapter 2, the advection-dispersion
equation was modified to model microbial transport and fate through each layer of porous
media (either saturated soil or unsaturated soil). The transport models give the final
number or concentration of microorganisms in the exposure media of groundwater well
water. The flowchart in Figure 3-2 shows how the three major components work to
39
complete the exposure assessment.
Figure 3-1 Groundwater exposure model
40
Figure 3-2 Flow chart for model development
The transport model through saturated porous media was developed using the
advection-dispersion equation incorporating both pathogen decay and adsorption to soil.
The exposure model was populated with microorganism-specific parameters, such as
occurrence concentrations in biosolids, decay rates in water, and soil-water partitioning
constants, and provides a time-dependent, microbial concentration profile as a function of
41
distance. The model considers microbial transport and fate in both saturated and
unsaturated soil.
The governing equation for microbial transport through saturated soil, both for
vertical wetting zone transport and horizontal groundwater flow, is the one-dimensional
advection-dispersion model with an instantaneous source, including effects of adsorption
to soil and first-order inactivation of pathogens (Bedient, et al., 1997) (Equation 3-1).
2
2x x
C C CD v C R
x x t
(3-1)
where Dx is coefficient of hydrodynamic dispersion (cm2/h) (= αvx, where α is
dispersivity), vx is the average seepage velocity (cm/h), λ is the first order inactivation
rate (1/h), and R is the retardation factor. The retardation factor is defined as 1+(ρb/n)Kd,
where ρb is the bulk dry mass density (g/cm3), n is porosity, and Kd is the equilibrium
distribution coefficient (cm3/g).
A filtration mechanism was included in the model to capture pathogen removal
due to physical straining and other filtration processes, which are particularly relevant for
larger microbes, such as bacteria and protozoa. The filtration removal is determined by
the coefficient (kstr) and the distance over which straining occurs (L). The preponderance
of straining occurs near the inlet, especially for the first 1 cm length (Foppen et al. 2007).
Equation 3-3 was used to predict the concentration of strained microorganisms (Tufenkji
2007).
0
exp( )str
kCL
C v
(3-3)
where C is the effluent concentration and C0 is the influent concentration, kstr is
the straining coefficient, which is estimated using a correlation based on the
42
microorganism and grain size ratio (Bradford, Simunek et al., 2003) , v is the interstitial
microbe velocity, and L is transport distance, which was assumed to be 1 cm in the model.
The fraction of strained microorganisms is calculated as (1-C/C0).
3.2.2 Field monitoring data
Lysimeters and a portable rainfall simulator were used to monitor control sites to
evaluate the leaching and ponding of viral contaminants following land application of
biosolids (Wong, Harrigan, et al., 2012). Mesophilic anaerobic digested (MAD) biosolids
were applied on sandy-loam soil (Fine-loamy, mixed, semiactive, mesic Typic
Hapludalfs). Portable rainfall simulators were used to apply water on a semi-continuous
basis to minimize surface ponding. Six large containment lysimeters (numbered
lysimeters L1 through L6) were used for leaching studies. Leachate samples were
collected in 2008 from 3 lysimeters (L1 to L3) and in 2009 from another 3 lysimeters (L4
to L6) and were analyzed for anionic tracer (chloride), microbial tracer (P-22
bacteriophage), adenoviruses and somatic phage (2009 study only).
3.2.3 Model validation
The subsurface fate and transport model was applied to predict microbial
concentrations in the effluent of the field lysimeters, and to estimate the field parameters
that best characterize subsurface transport. Based on the total amount of water applied per
day, the inputs of rainfall rate and duration were approximated for the model. The inputs
describing the site characteristics and application events are shown in Table A-5 in
Appendix. Due to the heterogeneous field conditions, hydraulic parameters, including
pore water velocity and dispersivity, were treated as fitting parameters for each of the six
lysimeters respectively, and used to predict effluent concentrations from the six
43
lysimeters. The outputs of microbial concentration profiles were examined, compared to
the observed results, and used to estimate model parameters (Table 3-1). Breakthrough
curves for chloride were used to adjust the hydraulic parameters, including pore water
velocity and dispersivity; the breakthrough curves for P-22 were used to adjust the
microbial release fraction, decay rate, and retardation factor. Since no adenovirus or
somatic phage was recovered from the leachate samples, adenovirus and somatic phage
concentrations were predicted using the fitted hydraulic parameters to confirm that
predicted concentrations were under the experimental detection limit. The Solver
function in Microsoft Excel was used to estimate model parameters by minimizing the
sum of the squared deviations between the modeled and observed concentration values.
44
Table 3-1 Summary of model validation results
Tracer or
microorg
anism
Approach Results
Chloride
Assumed
retardation factor of
1, decay rate of 0;
Adjusted hydraulic
parameters (pore
water velocity and
dispersivity) to fit
breakthrough
curves for chloride
in six lysimeters
individually.
Pore water
velocity
(cm/h)
Dispersivity
(cm)
Fraction not
captured by
lysimeter1
L1 4.34 95.8 0.287
L2 2.90 63.8 0.258
L3 6.56 27.9 0.674
L4 4.59 99.1 0.007
L5 2.91 105.1 0.258
L6 0.82 93.6 0.488
P-22 Used fitted
hydraulic
parameters for each
lysimeter; Adjusted
retardation factor,
recovery fraction
(restricting values
to less than the
chloride recovery)
and decay rate to fit
breakthrough
curves for P-22.
Retardat
ion
factor
C
aptured
fraction2
R
elease
fraction3
Decay
rate
(log/hr)
L2 3.46 0.273 0.368 0.011
L5 0.55 0.049
0.066 0.0032
L6 0.23 0.512 1 0
Adenovir
us
Predicted effluent
concentration using
published microbial
parameters and
fitted hydraulic
parameters.
Averaged predicted concentration in six lysimeters
is 3.53×10-2
, 4.61×10-2
, 2.95×10-2
, 3.21×10-3
,
4.03×10-3
, and 4.88×10-3
adenoviruses per ml.
Somatic
phage
Predicted effluent
concentration using
published microbial
parameters and
fitted hydraulic
parameters.
Averaged predicted concentration in three
lysimeters (L4, L5, and L6) a 7.93×10-6
, 9.96×10-6
,
and 1.21×10-5
phages per ml.
Note: 1 Fraction not captured by lysimeter was calculated from (1-recovery), where recovery is
measured mass recovery of chloride. 2 Captured fraction is defined as the overall percentage of P-22 leached out from the soil
media. It was fitted by Solver with an upper bound restricted to be the same as the
measured chloride mass recovery. 3 Release fraction is defined by the ratio of the overall captured fraction to the maximum
45
possible recovery fraction (assumed to be the same as the upper bound of chloride
recovery from field data)
3.3 Results
3.3.1 Anionic tracer (chloride) concentrations
The hydraulic parameters, including pore water velocity (cm/hour) and
dispersivity (cm) were adjusted to fit observed chloride breakthrough curves in six
lysimeters (Figure 3-3). A fraction not captured by lysimeter was assumed based on the
measured recovery percentage of chloride (Wong et al., 2010), retardation was assumed
to be 1, and the decay rate was set to 0 because chloride is a conservative tracer. The
results of fitted pore water velocity and dispersivity values are included in Table 3-1.
46
Figure 3-3 Predicted and observed breakthrough curves for chloride
Among all six studied lysimeters, the fitted velocity values ranged from 0.82 to
6.56 cm/hour. The predicted pore water velocities were higher than measured velocities,
0
10
20
30
40
0 2 4
chlo
rid
e c
on
cen
trat
ion
(p
pm
)
Pore volume
L1
Generated by SMART Biosolids
Observed by Wong et al., 2010
0
10
20
30
40
0 2 4
chlo
rid
e c
on
cen
trat
ion
(p
pm
)
Pore volume
L2
Generated by SMART Biosolids
Observed by Wong et al., 2010
0
10
20
30
40
0 2 4
chlo
rid
e c
on
cen
trat
ion
(p
pm
)
Pore volume
L3
Generated by SMART Biosolids
Observed by Wong et al., 2010
0
10
20
30
40
0 2 4ch
lori
de
co
nce
ntr
atio
n (
pp
m)
Pore volume
L4
Generated by SMART Biosolids
Observed by Wong et al., 2010
0
10
20
30
40
0 2 4
chlo
rid
e c
on
cen
trat
ion
(p
pm
)
Pore volume
L5
Generated by SMART Biosolids
Observed by Wong et al., 2010
0
10
20
30
40
0 2 4
chlo
rid
e c
on
cen
trat
ion
(p
pm
)
Pore volume
L6
Generated by SMART Biosolids
Observed by Wong et al., 2010
47
which were calculated from measured surface infiltration rates using an infiltrometer
inserted into the upper soil layer at the end of the experiments (Figure 3-4). Generally,
infiltration rates are highest at the beginning of a wetting event, then decrease over time
to a constant lower rate and achieve the lowest values at the end (Williams et al., 1998),
which is consistent with the observed data. Meanwhile, it has been widely reported that
macropores and fractures, such as burrows and plant root holes, create preferential flow
paths, thereby enabling rapid downward transport of microbes from the contamination
source to the water table (Duan et al., 2010; Weiler, 2005; Williams et al., 1998). Plant
roots in the soil beneath the six lysimeters, with a rotation of grass or corn on the surface,
may increase the infiltration by increasing the hydraulic conductivity of the soil, resulting
in a faster averaged velocity. In addition, the predicted velocity is the averaged value
through the entire lysimeter, which may be higher than the velocity in the upper soil layer.
The infiltration rates within each lysimeter vary by a factor around 7 from different sites
in the same field. However, there is a strong relationship between the fitted and measured
infiltration rates with the ratio of fitted and measured rates varying roughly from to 2 to 3.
The measured infiltration rate may underestimate bacterial transport rates, while the fitted
velocity provides a more protective estimation.
48
Figure 3-4 Comparison of the fitted velocity and measured velocity (calculated velocity
from field measured infiltration rate divided by porosity) in six lysimeters
The fitted dispersivities in six lysimeters range from 27.86 to 105.08 cm with
travel distance of 240 cm. These values are consistent with reported dispersivities from
literatures (Jaynes, 1991; van Wesenbeeck, et al., 1991; Xu and Eckstein, 1995;
Vanderborght, 2007). It is reported that the range of dispersivities is from 3.1 cm to 195.3
cm for 100- to 200-cm transport in field study (Jaynes, 1991; Wilson et al., 1998). It is
found that the dispersivity is sensitive to the scale of the experiments (Xu et al., 1995;
Vanderborght, 2007). Xu et al. (1995) used the weighted least-squared method in analysis
of the scale effect on dispersivity based on data from the large field experiments, and
proposed a positive relationship between the field scale and the dispersivity
(α=0.83(log10L)2.414
, where is dispersivity, and L is travel distance). Vanderborght et al.
(2006) derived a database of dispersivity values from leaching studies in soils and found
dispersivities increased with increasing transport distance and higher dispersivity
0
1
2
3
4
5
6
7
L1 L2 L3 L4 L5 L6
fitted velocity (cm/h)
measured velocity (cm/h)
49
obtained from field studies than column study. With a large travel distance of 240 cm in
the lysimeter field study, the dispersivities would be higher than the values observed in
laboratory column-scale study or field-scale study with smaller travel distances. It is also
observed that the microorganisms had comparable but slightly higher dispersivities than
chemical tracer (Sinton et al., 2010). Sinton et al. (2010) observed dispersivity of 1.0-1.9
cm for E-coli and 0.8 cm for bromide during a 3-m transport in lab. It should be noted
when using the dispersivity fitted by anionic tracer for modeling microbial transport.
3.3.2 Microbial tracer (P-22) concentrations
P-22 was only detected in lysimeters L2, L5 and L6. Using the hydraulic
parameters fitted for these three lysimeters from chloride tracer tests, the retardation
factor, release fraction and decay rate were adjusted to fit the observed breakthrough
curves for P-22. The fitted breakthrough curves are compared to observed data (Figure 3-
5). The results of fitted microbial parameters are presented in Table 3-1.
50
Figure 3-5 Predicted and observed breakthrough curve for P-22
Peak chloride breakthrough occurred at approximately 0.3 pore volumes in each
lysimeter, whereas peak P-22 breakthrough varied among the lysimeters (<0.1, 0.3 and
0.7 pore volumes) (Wong et al., 2010). The difference between chloride and P-22
breakthrough curves may be due to sorption, which can be described by the different
retardation factors for P-22 in lysimeters L2, L5 and L6. Using the following definition of
retardation (R=1+(ρb/n)Kd), and a reported Kd value (a partitioning coefficient between
soil and groundwater) of 0.557 cm3/g for enterovirus (Lyon et al., 2003) corresponds to a
retardation factor (R) of 3.41. This value is close to the observed retardation of P-22 in
0.00E+00
2.00E-05
4.00E-05
6.00E-05
8.00E-05
1.00E-04
1.20E-04
1.40E-04
1.60E-04
1.80E-04
2.00E-04
-0.5 0.5 1.5 2.5
C/C
0
Pore volume
L2
0.00E+00
2.00E-05
4.00E-05
6.00E-05
8.00E-05
1.00E-04
1.20E-04
1.40E-04
1.60E-04
1.80E-04
2.00E-04
0 0.5 1 1.5 2 2.5
C/C
0
Pore volume
L5
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
2.50E-03
3.00E-03
3.50E-03
4.00E-03
4.50E-03
5.00E-03
0 0.5 1 1.5 2 2.5
C/C
0
Pore volume
L6
51
L2 (R=3.46), but it is higher than the retardation factors in L5 and L6 (R=0.55 in L5 and
0.23 in L6, respectively). This discrepancy may be due to preferential flow paths, which
may limit the exposure of P-22 to sorptive soil surfaces (Kamra et al., 2001).
The decay rates directly fitted from the breakthrough curves are 0.011, 0.0032 and
0 log/hr in L2, L5, and L6, respectively, resulting in an average value of 0.0047 log/hr for
the die-off rate. These fitted decay rates are consistent with the reported decay rates for
coliphage (0 to 0.0042 log/hr in groundwater with temperature between 0 to 10 °C) (John
and Rose, 2005).
It has been reported that less than 8% of coliphage leaches out of the biosolids-
soil matrix (Chetochine et al., 2006). However, the release fractions in this study fitted to
0.368, 0.066, and 1 in L2, L5, and L6, respectively. The reason for this discrepancy is
that in the lysimeter study, the biosolids were spiked with P-22 bacteriophage
immediately before the biosolids application, so the spiked liquid may not have the
opportunity to bind as tightly to the solid phase, resulting in a higher fitted release
fraction.
3.3.3 Adenovirus and somatic phage concentrations
The subsurface fate and transport model was used to predict adenovirus and
somatic phage concentrations using the adjusted hydraulic parameter values fitted from
this study. The retardation factors, release fractions, and decay rates specific to each of
these two viruses were input as default values in the model (Enriquez, 1995; Kamra et al.,
2001; John et al., 2005; Chetochine et al., 2006; Xagoraraki, 2010). Average predicted
adenovirus concentrations in six lysimeters were 0.035, 0.046, 0.030, 0.0032, 0.0040, and
0.0049 adenoviruses per ml, which are all below the 0.1 virus per ml detection limit of
52
qPCR reaction. The averaged values of predicted somatic phage concentration in three
lysimeters (L4, L5, and L6) were 7.93×10-6
, 9.96×10-6
, and 1.21×10-5
phages per ml, all
of which are far below the 5 PFU/100ml detection limit of the double layer agar method
(Grabow et al, 1986). Thus the predicted results are consistent with the observed results
of no adenovirus or somatic phage detected in the effluent of the lysimeters (Wong et al.,
2010).
3.4 Discussion
Results from tracer studies conducted in large containment lysimeters
representative of field conditions in agricultural cropping systems are presented in this
work. Since the variability of soil physical and chemical properties may produce different
flow patterns and different hydraulic properties, six sets of hydraulic parameters (pore
water velocity and dispersivity) were fitted using chemical and microbial tracer
breakthrough data for the six lysimeters. As presented in the Results section, the fitted
hydraulic parameters (both velocity and dispersivity) for an anionic tracer were higher
than measured values reported from previous field studies (Figure 3-4). Preferential
paths in the soil of the lysimeters may be responsible for this discrepancy. This may be
due to the plant roots in the soil, which could provide rapid flow downward to the water
table. The water flow pattern in different lysimeters also affected microbial fate and
transport behavior (including sorption or attachment to soil, and decay).
Fast preferential flow, as indicated by observed high velocities and dispersivities,
may contribute to the low retardation for the microbial tracer P-22. The fitted decay rates
for P-22 differed in each of the three lysimeters due to the different breakthrough
conditions (time, velocity, etc). The peak breakthrough concentrations of P-22 in three
53
lysimeters were 1.21×10-4
(C/C0 in L2), 8.53×10-4
(C/C0 in L5), and 1.02×10-3
(C/C0 in
L6). The higher concentrations in L5 and L6 may be explained by the preferential flow
produced by the heterogeneity of the soil (Table 3-2). This assumption was confirmed by
the quick breakthrough time (0.15 PV for L5 and 0.06 PV for L6 as opposed to 0.5 PV in
L2). Time to breakthrough would also affect the overall decay rate in each lysimeter.
Since the travel is fast, microbial decay is fitted to 0 log/hr in L6. With the longer
transport time and more contact with soil in L2 (168 hour to breakthrough), the fitted
decay rate was observed to be as high as 0.011 log/hr.
There is considerable evidence that the greater the number of indicator organisms
in water, the greater the number of pathogens (National Research Council, 2004). The
indicator to pathogen ratios can be used to determine the pathogen concentrations based
on indicator concentrations, especially when the pathogen concentrations are below
detection limit. In order to interpret the indicator concentration in the effluent from
lysimeters and its prediction of pathogen concentration, we compared the ratio of
predicted indicator (P-22) and predicted pathogen (adenovirus) concentrations in the
effluent from three lysimeters (L2, L5 and L6) (Table 3-2) to the indicator to pathogen
ratios in the biosolids. The ratios in the effluents of L5 and L6 differ more from the
original indicator to pathogen ratio in the biosolids with large uncertainties (average
ratios in effluents of 4.79×104 and 1.42×10
5 in L5 and L6, respectively, compared to
original ratios in the biosolids of 1.5×102 in L5 and 2.95×10
3 in L6). One possible
explanation is that P-22 had an early breakthrough time (less than 0.3 pore volume) in
both L5 and L6 and the discrepancy of breakthrough time between P-22 and adenoviruses
can be explained by the retardation factor, which is determined by the sorption between
54
soil and water (Figure 3-6 and Table 3-2). The retardation factors used for adenovirus in
the three lysimeters is a default of 3.17, while fitted retardations are 3.46 in L2, 0.55 in
L5, and 0.23 in L6 for P-22. The indicator and pathogen has a comparable retardation in
L2, and the ratio in L2 is more consistent during the whole breakthrough process as
comparing to the other two lysimeters. The ratios for the total mass in the effluent from
all three lysimeters are comparable (the same magnitude or one magnitude higher) to the
original ratios in the biosolids.
As an indicator, P-22 performs successfully for prediction of pathogens in the
ratios of the total mass since it has a similar decay rates as adenovirus (Table 3-2).
However, P-22 might fail to predict pathogens during the breakthrough process if the
retardations of indicators and pathogens differ from each other.
55
(a)
(b)
(c)
Figure 3-6 Comparison of predicted concentration breakthrough curve for P-22 and
adenovirus
0.E+00
1.E-01
2.E-01
3.E-01
4.E-01
5.E-01
6.E-01
7.E-01
8.E-01
0.E+00
2.E+02
4.E+02
6.E+02
8.E+02
1.E+03
1.E+03
0 0.5 1 1.5 2 2.5
ad
en
ov
iru
s co
nce
ntr
ati
on
P-2
2 c
on
cen
tra
tio
n
Pore volume
L2
P-22
adenovirus
0.E+00
1.E-02
2.E-02
3.E-02
4.E-02
5.E-02
6.E-02
0.E+00
5.E+00
1.E+01
2.E+01
2.E+01
3.E+01
3.E+01
4.E+01
4.E+01
0 0.5 1 1.5 2 2.5
ad
en
ov
iru
s co
nce
ntr
ati
on
P-2
2 c
on
cen
tra
tio
n
Pore volume
L5
P-22
adenovirus
0.E+00
5.E-03
1.E-02
2.E-02
2.E-02
3.E-02
3.E-02
4.E-02
0.E+00
5.E+01
1.E+02
2.E+02
2.E+02
3.E+02
3.E+02
4.E+02
4.E+02
5.E+02
5.E+02
0 0.5 1 1.5 2 2.5
ad
en
ov
iru
s co
nce
ntr
ati
on
P-2
2 c
on
cen
tra
tio
n
Pore volume
L6
P-22
adenovirus
56
Table 3-2 Microbial breakthrough information and indicator:pathogen ratios in L2, L5 and L6
Peak
concentration
(C/C0)
Peak time (pore
volume)
Decay rate
(log/hr)
Retardation
factors
Ratio of P-
22:adenovir
us
concentrati
on in the
biosolids
Ratio of P-
22:adenovir
us total
mass in the
effluent
p-22 adenovi
rus
p-22 adenov
irus
p-22 adenovi
rus
p-22 adenov
irus
L2 1.21×10-4
5.65×10-5
0.5 0.7 1.1×10-2
1.75×10-3 a
3.46 3.17 b
7.14×102 7.82×10
2
L5 8.53×10-4
4.85×10-5
0.15 0.7 3.2×10-3
1.75×10-3 a
0.55 3.17 b 3.79×10
2 1.50×10
2
L6 1.02×10-3
2.83×10-5
0.06 0.47 0 1.75×10-3 a
0.23 3.17 b 3.79×10
2 2.95×10
3
a Values from literature (source: Enriquez, 1995)
b Values from SMART biosolids model (Galada et al., 2012a)
c Values displayed are averaged values with minimum and maximum estimations.
57
From the observed and predicted results of this lysimeter field study, there was a
3 to 6-log removal of indicators (P-22 bacteriophage and somatic phage), and a 7-log
removal of viruses (adenoviruses) achieved by transport through a 2.4-m-lysimeter. The
average removal rate was 1.8 log/m for indicators, which is consistent with the reported
removal rates for soil (larger than 1 log/m for most soil types) (Pang, 2009). It is also
reported that for the same media, removal of viruses is higher than removal of phage
species (Woessner et al., 2011). So the removal rate of 3 log/m for viruses is a reasonable
estimate. The results of this study contribute the removal of viruses and phages by sandy-
loam soil under field conditions.
58
CHAPTER 4. MODEL INTEGRATION - SPREADSHEET ENVIRONMENT
4.1 Introduction
Due to the complexity of risk assessments, models tend to be dense and difficult
for users to follow and modify in order to meet their needs. A spreadsheet-based tool,
named the Spreadsheet Microbial Assessment of Risk: Tool for Biosolids (SMART
Biosolids), has been developed for quantitative microbial risk assessment of land-applied
biosolids, which is intended to address these challenges. The model combines
spreadsheets with add-in visual basic macros in a rational and supportable manner.
Spreadsheets serve as a familiar interface for an archive of relevant inputs for parameter
values and references. The exposure model is also encoded in the spreadsheet, which
allows users to trace back computations through the model and modify parameters if
necessary. Add-in macros are used to implement a nested sampling routine that calls the
exposure model encoded in the spreadsheet many times to calculate values for different
pathogens and to perform a Monte Carlo uncertainty analysis. An example application
finds that adenovirus is the pathogen presenting the highest risk by all pathways.
However, uncertainties are large indicating that additional information on the fate and
transport of adenovirus in groundwater would be helpful. The SMART Biosolids model
may be useful for informing a number of decisions. Regulators and land application
program managers may be able to use the model to review different sites and determine
which sites are most appropriate for land application. Researchers may use the model to
integrate information and identify key gaps in knowledge warranting future research.
59
4.2 Methods
This chapter describes a computational model, named The Spreadsheet Microbial
Assessment of Risk: Tool for Biosolids (SMART Biosolids), including its spreadsheet
interface and add-in macros. SMART Biosolids model estimates risk associated with
exposure to pathogens from land-applied biosolids through five pathways (Figure 4-1).
These five pathways were identified through previous research efforts that developed a
framework for microbial risk assessment from land applied biosolids (Colford et al., 2003;
Eisenberg et al., 2004, 2006, and 2008). SMART Biosolids model assesses risk to highly
exposed individuals, such as residents whose homes border land application sites. This is
in keeping with the National Research Council recommendation that biosolids risk
assessments should focus on highly exposed individuals (NRC, 2002). The environmental
fate and transport models associated with each of the exposure pathways are computed in
Microsoft Excel (Galada, Gurian et al., 2012a). Each of the exposure pathway models is
described briefly below.
Inhalation of aerosols from land application sites is modeled by superposition of
Gaussian plume dispersion models from different locations on a grid representing the
field where the land application is taking place (Sehmel, 1980; Brooks, Gerba et al., 2004;
Low, Paez-Rubio et al., 2007). Consumption of groundwater affected by land-applied
biosolids is modeled by first using a Green-Ampt model to determine the depth of the
wetting front associated with any wet weather events (Swartzendruber, 1974; Salvucci
and Entekhabi, 1994; Zhang, 2009), and then a series of one-dimensional advection-
dispersion models (Bedient, Rifai et al., 1997; Faulkner, Lyon et al., 2002) are use to
describe transport through the soil. A microbial transport model for saturated media is
60
used to describe vertical transport of microbes to the end of the wetting front, and then a
transport model for unsaturated media is used to describe vertical transport to the water
table (this step is skipped if the wetting front extends to the water table). Finally the
saturated media transport model is used to describe horizontally from the field to a down
gradient well. Contributions from different locations in the field are superimposed to
obtain the net concentrations at the well. Direct ingestion of biosolids-amended soils is
modeled by allowing for first order decay of applied microorganisms followed by the use
of standard exposure factors for incidental ingestion of soil (U.S. EPA, 1997).
Consumption of water contaminated by runoff from a land application site is modeled
first estimating runoff from wet weather events using the Green-Ampt infiltration model
(Swartzendruber, 1974; Salvucci and Entekhabi, 1994; Zhang, 2009). Then the Revised
Universal Soil Loss Equation (RUSLE, 2008) is revised and used in a finite difference
approach to track net eroded vs. deposited soil (with associated biosolids) over a one-
dimensional flow path. Transport of free microbes (those not associated with soil
particles) is tracked separately. Both free and soil-associated microbes are assumed to
runoff to a pond with human exposure occurring by full-contact recreation in the pond.
Ingestion of plants impacted by land-applied biosolids is modeled by assuming that the
runoff goes into an adjoining field with a portion retained by the leaves of a lettuce plant.
First order decay is modeled between the time of contamination and consumption. For all
models standard exposure factors (U.S. EPA, 1997) and literature dose-response models
are used to calculate risks based on the environmental concentrations estimated from the
different pathways models.
61
Figure 4-1 Exposure pathways considered
4.2.1 Software availability
Software name: Spreadsheet Microbial Assessment of Risk: Tool for Biosolids
(SMART Biosolids)
Year first available: 2011
Software required: Microsoft Excel
Program languages: Spreadsheet and Visual Basic
Program format: Microsoft Excel macro-enabled workbook (.xlsm file)
Program size: 3.8 MB
Availability: The manual for the SMART Biosolids model and the CD including
the spreadsheet tool are available in Galada, Gurian et al. (2012a). A website link to the
model will be provided by the Water Environment Research Foundation.
62
4.2.2 Model development in spreadsheet
The environmental fate and transport models associated with each of the exposure
pathways are computed in Microsoft Excel (Galada, Gurian et al., 2012a). A spreadsheet
interface is used, which provides a familiar interface for entering inputs and accessing
outputs while providing access to a range of post-processing tools including chart-
building capabilities. The inputs required in the model include site-specific parameters
(such as parameters associated with application events, climate and soil characteristics),
pathogen-specific fate and transport parameters, as well as those for uncertainty analysis
(such as the iteration numbers, and option to restore nominal parameter values after an
uncertainty analysis). The outputs include estimates and uncertainties of expected
concentrations of microbes in air, soil, surface and groundwater resulting from biosolids
applications, and expected probability of infection by these microbes for residents of
nearby properties and workers applying the biosolids.
The use of a spreadsheet allows the advanced user to trace the computations used
and modify parameters and even change mathematical algorithms as desired.
Spreadsheets provide a useful storage for the information collection, including
representative data and references. Citations with author names and dates are marked
next to the data in the sheet, and full references are listed at bottom of each sheet. A three
tiered color coding system is used to differentiate between required user inputs (blue
cells), defaults values (green cells), and modeling results (yellow cells). User-friendliness
was tested by professionals in the biosolids field who had no background in statistics or
risk assessment.
63
4.2.3 Use of Visual Basic macros
The exposure models are complex and require that a substantial amount of
information be kept in working memory in order to have these computations available for
the user in the spreadsheet environment. For example, superposition is used to estimate
the cumulative effect of spatially dispersed sources for the air and groundwater pathways
and a finite difference approach is used to track sediment erosion and deposition over a
flow path for the surface water model. The same exposure model is used multiple times
with different parameters and different possible input values (that is, Monte Carlo
uncertainty analysis is performed for each organism). Unfortunately commercially
available Monte Carlo add-ins (e.g. @Risk, Crystal Ball) do not readily support nested
sampling. Thus, it would be necessary to conduct separate analyses for each
microorganism (i.e. keep separate copies of the exposure model for each microorganism
in the spreadsheet). The approach taken here is to use Visual Basic macros to execute a
loop that cycles through all 28 pathogens of concern. This loop is nested inside a Monte
Carlo uncertainty analysis (Figure 4-2). This cuts down on the amount of material in
working memory by limiting the spreadsheet coding to a single copy of the exposure
model. The Visual Basic code is readily viewable and can be edited by the advanced user.
64
Figure 4-2 Flowchart of the SMART Biosolids model
4.3 Results
Example results of the SMART Biosolids model are presented to demonstrate
how the model enables comparisons of risk across pathogens, comparisons across the five
exposure pathways, and the identification of key uncertainties. Model input parameters
were developed after consideration of typical biosolids land-applications observed in
65
Michigan.
The model required 80 minutes on a personal computer to finish a 1000-iteration
Monte Carlo uncertainty analysis. Details on important model inputs are provided in
Table 4-1. Figure 4-3 shows the estimated risks (with uncertainties) from different
organisms across the five exposure pathways. Adenovirus has the highest nominal risk
estimates from five exposure pathways, exceeding the next highest risk due to
Cryptosporidium and Giardia lamblia by almost 2 orders of magnitude. While these
results are based on many assumptions, including the use of idealized, homogeneous
transport models, when interpreted cautiously the results can help prioritize among
different risks and identify future research needs. In this case the model uncertainties are
large and noteworthy. Further research could be directed towards studying the occurrence
and transport of adenoviruses, which have both a high nominal risk estimate and very
substantial uncertainty.
66
Table 4-1 Site-specific input parameters
Parameter Value Unit
Time of start of rain after biosolids
application
0 days
Soil texture class sandy_loam -
Biosolids application rate 2.57 dry tons biosoilds/acre
Water table depth 3 ft
Distance to well 100 Ft
Hydraulic gradient 0.04 -
Rainfall rate 7.3 cm/h
Rainfall duration time 1 h
Figure 4-3 Columns show representative nominal risks across pathogens for exposure
through groundwater. Error bars represent the 5th
and 95th
percentiles.
4.4 Discussion
This chapter describes a tool, which links quantitative microbial risk assessment
with microbial fate and transport modeling. The spreadsheet format provides a flexible
1.E-151.E-141.E-131.E-121.E-111.E-101.E-091.E-081.E-071.E-061.E-051.E-041.E-031.E-021.E-01
1.E+00Air pathway
IncidentalIngestion ofSoil
Surface waterpathway
Groundwaterpathway
Ingestion ofcontaminatedcrops
67
and familiar interface and serves as an archive for parameters with associated references.
Add-in macros are used to perform the many repeated computations required to perform
Monte Carlo uncertainty analysis for multiple pathogens. At the same time, the user
should always be mindful that there are several inherent limitations of spreadsheets
(Pitblado, 1994): Spreadsheet tools are easy to modify, but leave no trail to identify
changes; formulas are expressed in column and row labels and need to be located to
understand; simple spreadsheet errors can compromise parts of the spreadsheet model. In
order to avoid these problems, there is a “restore” option that can reset the data to default
inputs.
The model is able to quantify risks for six pathogens: Giardia lamblia,
Cryptosporidium, Salmonella, Shigella, enterovirus, and adenovirus. The occurrence of
other pathogens in biosolids was either too low to be reliably quantified or lacking
altogether. Part of the objective of the tool development was to integrate available
knowledge and in the process identify gaps in existing knowledge for which future
research is warranted. Thus, quantified occurrence levels of additional pathogens can
extend the model and allow for risk estimates to be obtained for a broader set of
pathogens.
68
CHAPTER 5. MODEL APPLICATION – UNCERTAINTY ANALYSIS AND
SENSITIVITY ANALYSIS
5.1 Introduction
Environmental risk assessment is complex, requiring many inputs, many of which
may be subject to significant uncertainties. Uncertainty analysis is essential for
assessment of complex systems in order to determine uncertainty in the results due to
uncertainties in model inputs (Helton et al., 2006). Recently, uncertainty analysis in the
environmental field has envolved from qualitative assessment (high, moderate, low) to
quantitative analysis (Metzger et al., 1998). One effective and widely applied technique is
the Monte Carlo analysis (Hammonds et al., 1994; Morgan and Henrion, 1990; Turner et
al., 1985; Seiler and Alvarez, 1996). A Monte Carlo analysis includes generation of
model inputs from probability distribution functions describing the range of plausible
parameter values, propagation of uncertainty through the analysis by computation of
model outputs based on the multiple sets of generated inputs, and presentation of the
results of the analysis (Helton, 1993; Helton et al., 2006).
The uncertainties in risk assessments may due to uncertainties in the input
parameters, such as the spatial variability of bacterial concentrations in applied manure,
and the hydraulic properties of the soil texture (Guber et al., 2011), and lack of
knowledge as to the appropriate structure of the transport and risk model. Sensitivity
analysis can be used to determine the contributions of individual uncertain inputs to the
uncertainty in analysis results (Helton et al., 2006; Frey and Patil, 2002; Cullen and Frey,
1999).
The microbial risk assessment tool for land-applied biosolids takes several crucial
elements into account, including occurrence values of pathogens in biosolids, the
69
potential routes of infection, the amount that humans would ingest, and the virulence of
the infectious agent (Galada et al., 2012b). The large uncertainties within all these factors
affect the confidence in the final prediction of risks. Uncertainty analysis is presented
here to show the range of possible output results. Sensitivity analysis is used to identify
the relative importance of different input uncertainties, and to prioritize additional data
collection or research.
5.2 Methods
The SMART Biosolids model is applied to typical site-specific conditions and
risks are computed for five pathways: consumption of water contaminated by runoff from
a land application site (surface water pathway), consumption of groundwater
contaminated by biosolids (groundwater pathway), inhalation of aerosols from land
application sites (air pathway), direct ingestion of biosolids-amended soils (soil pathway),
and ingestion of vegetables grown on biosolids-amended fields (vegetables pathway). As
quantitative occurrence information is available for only six pathogens, including
Salmonella, Shigella, adenovirus, enteroviruses, Cryptosporidium and Giardia lamblia,
these are the pathogens considered in this analysis. Risk assessment models are used to
calculate cumulative risk of illness over time for selected pathogens. An overview of the
SMART Biosolids model is provided in Chapter 4. The detailed exposure models and
descriptions of scenarios can be found in the manual for the SMART Biosolids model
(Galada, et al. 2012a).
5.2.1 Risk models
Risks of infection and illness depend on exposed dose, pathogen type, and
pathogen-specific dose-response models. Dose was calculated based on the predicted
70
environmental concentration from microbial transport and fate models (Equation 5-1).
Exponential dose-response models were used to calculate daily risk from a particular
pathogen (Haas, Rose et al., 1999) (Equation 5-2).
(5-1)
(5-2)
where Concdaily,i is environmental concentrations of the ith pathogen (number/L),
Exp is exposure rate (L/day), Dosedaily,i is the daily dose of the ith pathogen (number),
Riskdaily,i is daily risk of infection from the ith particular pathogen, r is the fraction of the
ingested microorganisms that survive to initiate infection. For pathogens that have Beta-
Poisson dose response models, the Taylor series approximation for low risks (ratio=α/β)
is used to correctly calculate risk (Teunis et al., 1996).
The risks calculated from Equation 5-2 represent the daily initial risks for the first
day of exposure. In order to examine the risk over a longer time, the cumulative risk is
calculated. For the air inhalation pathway, the cumulative risk over time was calculated
by incorporating risk from slinger application (Riskslinger) followed by disk incorporation
on the same day (Riskdisk).
(5-3)
For accidental ingestion of soil, the risk estimate for time t (Riskt), can be
calculated using risk at day 0 (Risk0) (in the model, day 0 here is actually 31 days after
biosolids application as the Part 503 regulation require a 30 day site access restriction)
and the pathogen decay rate (kdecay).
(5-4)
, ,daily i daily iD ose C onc Exp
,( )
,1
daily ir Dose
daily iRisk e
71
A geometric series can be formed based on the sum of these estimates for
different time periods (assuming that the risks are small).
∑ ( )
(5-5)
For large N, the exponential term tends to zero, which further simplifies the
equation.
( )
(5-6)
For ingestion of surface water, groundwater and contaminated vegetables
pathways, the risk in the model outputs is conditional given the occurrence of particular
rainfall events (Chapter 2; Teng et al., 2012). The risk at day 0 can be calculated by
adding the product of the probability of the rainfall occurrence (P2-yr, P5-yr, P10-yr, P25-yr,
P50-yr, P100-yr) and the conditional risk (the risk predicted given this particular rainfall)
(Risk2-yr, Risk5-yr, Risk10-yr, Risk25-yr, Risk50-yr, Risk100-yr). The precipitation data,
including intensity and duration, is available for return periods of 2, 5, 10, 25, 50, and
100 years (Equation 5-7). As derived above, the risk over time can be simplified to
Equation 5-8.
∑ ( ) (
) ( ) ( )
( )
(5-7)
(5-8)
5.2.2 Model inputs
Risks from groundwater are compared across organisms and compared to other
72
pathways under the same site-specific conditions. Tables 5-1 and 5-2 summarize the
default inputs for site-specific conditions. These inputs do not correspond to a particular
site but were developed after consideration of typical applications observed in Michigan
by Kumar et al. (2010). For ingestion of surface water, ingestion of contaminated
vegetable, and ingestion of groundwater, the exposure prediction is affected by the
intensity and duration of wet weather events. Table 5-2 shows the model inputs of rainfall
intensity and duration, which are based on the determination of the rainfalls producing
maximum infiltration and runoff depths (Chapter 2; Teng, 2010).
73
Table 5-1 Site-specific conditions
Parameter Value Unit
Time of start of rain after biosolids
application 0 days
Temperature 83 Fahrenheit
Soil texture class sandy_loam -
Area of application site 625 Acre
Slope of the plot 4.00 %
Application method Slinger and disk
incorporation None
Biosolids application rate 2.57 dry tons biosoilds/acre
Water Table Depth 3 ft
Distance to Well 100 Ft
Hydraulic Gradient 0.04 -
Presence of buffer strip between
field and ditch (VS) 1 1(Yes) or 0(No)
Length of buffer strip 33 Ft
Slope of buffer strip 4.00 %
Presence of channel after VS 0 1(Yes) or 0(No)
Presence of pond 1 1(Yes) or 0(No)
Distance of residential population
to field 250 Ft
Time of soil ingestion after
biosolids application 31 Days
Time for exposure to pond water
after biosolids application 0.0000001 Days
Consider resuspension for
occupational workers during
biosolids application
1 1(Yes) or 0(No)
Time of vegetable ingestion after
biosolids application 5 Days
Number of ingested vegetable
leaves 5
Computational Reporting
Threshold 1.00E-20
Lowest level of risk
reported
74
Table 5-2 Critical rainfall event information (see Chapter 2 for source)
Return
period
(year)
Probability of
occurrence in
one day
Surface water pathway and
contaminated vegetable pathway
Groundwater pathway
Rainfall
duration
(h)
Rainfall
intensity
(cm/h)
Runoff
produced
by
rainfalls
(cm)
Rainfall
duration
(h)
Rainfall
intensity
(cm/h)
Wetting
front
depth
produced
by
rainfalls
(cm)
2 1.37×10-3
0.25 7.62 0 0.5 5.18 9.74
5 5.48×10-4
0.25 9.51 0.25 1 4.26 14.55
10 2.74×10-4
1 4.92 0.65 1 4.92 14.55
25 1.10×10-4
0.5 8.99 1.35 1 5.36 14.55
50 5.48×10-5
0.5 10.08 1.90 2 2.84 16.16
100 2.74×10-5
1 7.3 2.59 2 3.20 22.14
Note: Rainfall values for maximum runoff and infiltration based on different soil
hydraulic property parameters for surface water (Pajian, 1987) and groundwater (Carsel
and Parrish, 1988) transport scenarios
5.2.3 Uncertainty analysis
The default best estimates and their associated probability distributions were
developed by 1000-iteration Monte Carlo uncertainty analysis performed using Microsoft
Office Excel’s add-in Visual Basic Macros to estimate the cumulative risks over time.
The model inputs used in the simulations are listed in Table A-6 in Appendix with
detailed information on distribution and resource. The inputs were divided into two
categories: microbial parameters (those that change for different types of microorganisms)
and soil parameters (those that change with different soil textures). There were also
several important parameters affected by both microorganism type and soil type, such as
effective dispersion factor, and retardation factor.
When available, parameter uncertainty distributions used standard deviations
from literature sources. In some cases, only the upper bound or lower bound was reported
75
for a parameter-of-concern without reporting any standard deviation. If the associated
percentile was available then the standard deviation was calculated based on the
assumption that the variable was normally distributed. If the associated percentile of the
upper bound or lower bound was not available, it was assumed that the reported upper
bound or lower bound corresponded to a 95th
or 5th
percentile respectively. If the upper
bound or lower bound was not available, the uncertainty factor was assumed to be 10.
5.3 Results
5.3.1 Uncertainty analysis
Figure 5-1 and Table 5-3 shows representative risks across pathogens from
exposure through five pathways.
76
Figure 5-1 Plots show cumulative risks over time for exposure through five pathways.
Error bars represent the 5th
and 95th
percentiles. Risks for adenovirus, Cryptosporidium,
enteroviruses, and Giardia lamblia are risks of minor illness cumulative over time; risks
for Salmonella and Shigella are risks of major illness cumulative over time
77
Table 5-3 Cumulative risks over time for exposure from five pathways with uncertainties
Air Surface
water
Soil Veg Groundwat
er
Adenovirus
4.13×106
(7.26×10-7
,
8.68×10-6
)
4.84×10-6
(NA,
5.24×10-4
)
2.36×10-2
(3.26×10-14
,
1)
1.38×10-6
(NA,
1.97×10-4
)
NA (NA,
2.43×10-3
)
Cryptosporidium NA
9.31×10-8
(NA,
8.09×10-6
)
4.52×10-4
(7.50×10-15
,
4.10×10-2
)
2.49×10-8
(NA,
3.55×10-6
)
NA (NA,
7.87×10-6
)
Enteroviruses
1.25×10-9
(2.50×10-10
,
1.00×10-9
)
1.52×10-10
(NA,
5.32×10-8
)
4.64×10-8
(NA,
1.23×10-4
)
2.32×10-11
(NA,
2.06×10-8
)
NA (NA,
8.19×10-10
)
Giardia lamblia NA
1.20×10-7
(NA,
1.29×10-5
)
6.99×10-5
(NA,
3.25×10-2
)
1.69×10-8
(NA,
3.49×10-6
)
NA (NA,
1.16×10-7
)
Salmonella spp.
5.67×10-13
(8.77×10-16
,
2.31×10-10
)
1.49×10-13
(NA,
5.07×10-10
)
2.88×10-12
(NA,
2.48×10-7
)
6.56×10-15
(NA,
1.34×10-10
)
NA (NA,
9.45×10-16
)
Shigella spp.
2.37×10-8
(2.39×10-9
,
8.69×10-8
)
2.90×10-8
(NA,
2.19×10-6
)
1.96×10-5
(NA,
8.67×10-3
)
3.22×10-9
(NA,
1.06×10-6
)
NA (NA,
1.14×10-7
)
Note: Risks for adenovirus, Cryptosporidium, enteroviruses, and Giardia lamblia are
risks of minor illness cumulative over time; risks for Salmonella and Shigella are risks of
major illness cumulative over time. Values displayed are averaged values with 5-95th
percentiles given in parentheses.
Adenovirus presents the greatest risks across different pathways. Cryptosporidium
and Giardia lamblia produce the next highest risk levels. The risks from these two
organisms differ by an order of magnitude or less by the same pathway. For the air
pathway, Cryptosporidium and Giardia lamblia present no risk, since protozoan
pathogens are not an issue in air due to their size (personal communication with Dr.
Charles P. Gerba, 2010). The other three pathogens, Shigella, enteroviruses, and
Salmonella, show lower risks.
The risks produced across pathways are ranked, in descending order, as soil,
78
surface water, vegetable ingestion, air, and groundwater. The soil pathway produced the
highest risks. All the other four pathways produce nominal risks lower than 1 ×10-5
. The
best estimates of risks from groundwater pathway are zeros, however, large uncertainties
are present due to the complexities of the exposure model, that includes many uncertain
input parameters. Looking at the upper bound of the risk estimates, the risk of minor
illness produced by adenoviruses is 2.43×10-3
. Although this is the estimation of
cumulative risk over time for one single application event, it exceeds the 1 in 10,000
benchmark associated with reported annual microbial risk from U.S. drinking water
supplies (Galada et al., 2012a). The risks of minor or major illness by the other five
pathogens do not exceed the 1 in 10,000 benchmark.
5.3.2 Sensitivity analysis
From the results of the Monte Carlo analysis (Figure 5-1), there are large
noteworthy uncertainties in the outputs. Input-Output correlations were examined to
identify the most important source of uncertainty for the predictions of risks for each of
the five pathways (Table 5-3). The inputs include microbial parameters for adenoviruses,
and soil parameters for sandy-loam soil. The outputs are the cumulative risks over time
from adenoviruses. Only results for risks from adenoviruses and sandy-loam soil are
presented here, but this approach can be used to identify the importance of inputs for
other pathogens or soil textures.
The spearman correlations were analyzed using SPSS, and the correlation is
determined to be significant at the 0.01 level (2 tailed) (Table 5-4). From the input-output
correlation table, it is found that four microbial inputs, microbial initial concentration in
biosolids, release fraction, microbial decay rate, and pathogen dose-response models, are
79
important for risks as expected. Initial concentration in biosolids is statistically significant
for inhalation risk (correlation=0.274). Microbial release parameter is only significant for
two pathways and the correlation coefficients are not very high (correlation=0.121 and
0.137 for surface water and vegetable pathways, respectively), as its effects can be
diluted by occurrence concentration. In the model, both occurrence concentration in
biosolids and release fraction affects the initial number of microorganisms entering
exposure transport. The decay rate is statistically significant except for the air pathway
(correlation=-0.979, -0.478, -0.300, -0.502, for soil, surface water, groundwater, and
vegetable pathways respectively). Dose-response models show statistically significant
correlations (correlation=0.941, 0.142, 0.124, and 0.168 for air, soil, surface water and
vegetable pathways, respectively) except for the groundwater pathways
(correlation=0.032). Saturated water content and residual water content only affect three
pathways, surface water, groundwater and vegetable ingestion and, none of the
correlations are significant. It is found that none of the inputs is statistically significant
for the risk produced by groundwater exposure pathway. It might be that because the
output of risks from groundwater pathway is very low, in many cases below the reporting
threshold of 10-20
, it is hard to identify the relationship between inputs and outputs.
80
Table 5-4 Input-output correlations for risks of illness cumulative over time from adenovirus
Description
Risk produced
by inhalation of
air
Risk produced
by incidental
ingestion of soil
Risk produced
by ingestion of
surface water
runoff
Risk produced
by ingestion of
groundwater
Risk produced
by ingestion of
contaminated
vegetables
Microbial initial
concentration in biosolids 0.274
a 0.052 0.005 0.043 0.012
Microbial release parameter 0.012 0.012 0.121a 0.038 0.137
a
Decay for microbes in
water/soil/air 0.003 -0.979
a -0.478
a -0.300
a -0.502
a
Pathogen ingestion- or
inhalation-related dose-
response models 0.941
a 0.142
a 0.124
a -0.032 0.168
a
Microbial radius N/A N/A N/A 0.053 N/A
Hydraulic conductivity N/A N/A N/A 0.050 N/A
Saturated water content N/A N/A 0.008 0.013 0.047
Residual water content N/A N/A -0.077 -0.002 0.011
Effective dispersion factor N/A N/A N/A 0.029 N/A
Retardation factor N/A N/A N/A -0.008 N/A
a Bold values represent the correlations are statistically significant.
81
Table 5-5 Ratios of pathogen:indicator in biosolids and in the environment
Pathogen:indicator ratio In the
biosolids
Air pathway Incidental
ingestion of
soil
Surface water
pathway
Groundwater
pathway
Ingestion of
contaminated
crops
Adenovirus: Coliphage 8.42×10-5
8.42×10-5
1.95×10-4
8.42×10-5
2.72×10-4a
9.64×10-5
E.coli 5.57×10-3
5.57×10-3
2.95×106 5.57×10
-3 4.11×10
7a 1.42×10
-1
Enterococci 1.39×10-3
1.39×10-3
4.84×103 1.39×10
-3 8.63×10
4a 1.57×10
-2
Fecal
coliforms
1.39×10-6
1.39×10-6
4.48×10-1
1.39×10-6
7.46a
1.07×10-5
Cryptosporidi
um:
Coliphage 1.34×10-4
1.34×10-4
3.52×10-4
1.34×10-4
3.58×10-5a
1.57×10-4
E.coli 8.86×10-3
8.86×10-3
5.33×106 8.86×10
-3 5.42×10
6a 2.31×10
-1
Enterococci 2.20×10-3
2.20×10-3
8.74×103 2.20×10
-3 1.14×10
4a 2.56×10
-2
Fecal
coliforms
2.20×10-6
2.20×10-6
8.08×10-1
2.20×10-6
9.82×10-1a
1.74×10-5
Shigella: Coliphage 2.15×10-5
2.15×10-5
6.93×10-6
2.15×10-5
3.72×10-6a
1.79×10-5
E.coli 1.42×10-3
1.42×10-3
1.05×105 1.42×10
-3 5.62×10
5a 2.64×10
-2
Enterococci 3.54×10-4
3.54×10-4
1.72×102 3.54×10
-4 1.18×10
3a 2.92×10
-3
Fecal
coliforms
3.54×10-7
3.54×10-7
1.59×10-2
3.54×10-7
1.02×10-1a
1.99×10-6
aWith default setbacks of 3-ft water table and 100-ft distance to well, zero concentrations were predicted for groundwater
pathway. In order to calculate the pathogen and indicator ratio, the setbacks were changed to 1-ft water table and 30-ft distance
to well.
82
5.3.2 Indicator and pathogen relationship
The ratios of pathogen concentrations and indicator concentrations are
presented to examine the pathogen and indicator relationship (Table 5-5). One
pathogen was selected for each of the pathogen types, adenovirus for viral pathogens,
Cryptosporidium for protozoan pathogens, and Shigella for bacterial pathogens. The
ratios in the five different exposure media can be compared to the original ratios in
biosolids. If the ratio in the biosolids and in the exposure media is consistent, it can be
concluded the indicator organism is predictive of the pathogen’s concentration. All of
the four indicators, coliphage, E.coli, enterococci, and fecal coliforms, performed as
good indicators for the air pathway and surface water pathway. For example, the
adenovirus: E.coli ratio was 5.57×10-3
originally in the biosolids. The ratios in air and
surface water pathway were both 5.57×10-3
. The next best performance was for
vegetable pathway. Adenovirus: E.coli ratio was 1.42×10-1
in vegetable pathway,
which indicates E.coli died off faster than adenovirus through this exposure scenario.
The indicators performed poorly for the soil ingestion and groundwater pathway.
Adenovirus: E.coli ratio increased to 2.95×106 and 4.11×10
7, which means adenovirus
was much more persistent than E.coli under these two scenarios. The reason can be
explained by that both soil and groundwater pathways need longer retention time than
air, surface water or vegetable pathway.
Since air pathway and surface water pathway give exactly same ratios as the
ratios initially in the biosolids, the abilities of prediction for different indicators can be
compared for the vegetable, soil and groundwater pathways. It is found that coliphage
were very predictive for all pathogen types. E.coli, Enterococci and fecal coliforms
were very poor indicators, especially for the soil and groundwater pathway. For
example, in groundwater pathway, adenovirus:coliphage ratio (2.72×10-4
) was an
83
order of magnitude higher than the ratio in biosolids (8.42×10-5
); adenovirus:E.coli
ratio was ten magnitudes (4.11×107) higher than the ratio in biosolids (5.57×10
-3);
adenovirus:Enterococci ratio (8.63×104) and adenovirus:fecal coliform ratio (7.46)
was 7 magnitudes higher than the ratio in biosolids (1.39×10-3
and 1.39×10-6
). The
reason is coliphage has similar decay rate as the pathogens, and E.coli, Enterococci
and fecal coliforms have an order of magnitude higher decay rates. Indicators fail to
indicate the pathogens’ concentrations when they decay faster than pathogens, which
is shown by a higher pathogen:indicator ratio in the environment than in the biosolids.
The failure is especially obvious for those pathways needing longer transport time
(such as the groundwater pathway) or with a time restriction before contact occurs
(such as the soil pathway).
5.4 Discussion
The results can be compared to literature values for annual probability of
infection due to biosolids land application (Viau, Bibby et al., 2011). Although this
chapter did not examine the annual risk, the predicted cumulative risk over time from
accidental direct ingestion is consistently with previous study: it is highest compared
to other pathways. Eisenberg et al. (2006) developed and demonstrated a microbial
risk assessment framework for biosolids-associated pathogens through direct
ingestion of biosolids-amended soils, inhalation of aerosols, and consumption of
groundwater. Since Eisenberg et al. (2006) used enteroviruses as the pathogen-of-
concern, their risk results are compared to the predicted enteroviruses risks in this
chapter (Table 5-6). The estimated single-event risks by Eisenberg et al. (2006) are
higher than estimates of cumulative risks in this chapter from all three pathways. One
of the reasons is Eisenberg et al. (2006) used higher occurrence numbers (log-normal
84
distribution with mean=1.13 and standard deviation=2.17, unit in PFU/g). The
occurrence information in SMART biosolids model (log-normal distribution with
mean=0.105 and standard deviation=0.2, unit in PFU/g) is more recent (Pepper et al.,
2010). Another important reason for the higher risk estimation from Eisenberg et al.’s
study is they used the rotaviruses dose-response models for risk estimates of
enteroviruses. It was reported the exponential parameters for inhalation dose-response
model are 0.31 for rotavirus and 0.025 for enteroviruses, and ingestion dose-response
parameters are 0.62 for rotavirus and 0.002 for enteroviruses (Regli et al., 1991;
Tanner et al., 2008; Brooks et al., 2005a, b; Haas et al., 1999). The discrepancy of the
risk estimates is also due to differences in the scenarios assumed in the two studies
(Table 5-6). For the soil pathway, half of the ingestion rate in Eisenberg et al. (2006)
was used in SMART Biosolids model, and 31-day restriction was also considered in
the SMART Biosolids model while Eisenberg et al. assumed immediate exposure. For
the air pathway, a large setback and lower wind speed was used in SMART Biosolids
model. For the groundwater pathway, the SMART Biosolids model does a more
through estimation by considering the effects of rainfall event. The predicted effect of
porous media for groundwater transport is much protective (longer transport distance
in unsaturated soil and with consideration of horizontal transport to well) than the
scenario assumed by Eisenberg et al. (2006).
85
Table 5-6 Comparison of enteroviruses risk estimates
Pathway Eisenberg et al. (2006) SMART Biosolids model
Risk estimates Scenarios Risk estimates Scenarios
Soil ingestion 7×10-4
(NA,
4.7×10-3
)
Ingestion rate
of 100 mg/day
immediately
1.92×10-8
(NA,
1.21×10-4
)
Ingestion rate
of 50 mg/day
on the 31st day
Air inhalation 7×10-5
(5.2×10-
5, 8.8×10
-5)
Wind speed
2m/s, and
distance is 30
m
1.25×10-
9(1.18×10
-10,
1.70×10-8
)
Wind speed is
0.8 ft/s, and
distance is 76 ft
Groundwater
ingestion
2×10-4
(NA,
4.2×10-3
)
Vertical
unsaturated soil
0.5 m, and
vertical
saturated soil 5
m
0(NA,
7.98×10-10
)
Different
scenarios
produced by
different
rainfall events.
The ranges are
0.08 to 0.2 m
for vertical
saturated soil,
0.71 to 0.83 m
for vertical
unsaturated
soil, and 30 m
of horizontal
saturated soil
Note: Values displayed are based on nominal input parameter values with 5-95th
percentiles of a Monte Carlo simulation given in parentheses. NA – not available as
value was below reporting threshold of 10-20
.
86
CHAPTER 6. CONCLUSION
This dissertation describes the development and application of a spreadsheet-
based quantitative microbial risk assessment framework for land-applied biosolids. The
contributions of this dissertation include: an approach to address the wet weather events
for QMRA using historical intensity-duration-frequency curves; a microbial subsurface
fate and transport model for exposure modeling of the groundwater pathway in QMRA
which includes the effects of wet weather events; a spreadsheet-based tool integrating the
most up-to-date data and analysis; and the application of the framework to compare the
risks across pathogens and pathways.
Chapter 2 presents a case study illustrating the use of widely available intensity-
duration-frequency curves to develop risk estimates for infiltration and runoff attributable
to wet-weather events. Infiltration and runoff for all storm events on a given frequency
curve are considered, and the maximum infiltration and runoff are associated with the
frequency for the specific curve. Because longer duration, less-intense storms tend to
have higher infiltration whereas shorter duration, more intense storms tend to have higher
runoff amounts, the critical runoff and infiltration events for a given return period will
often be produced by different storms on the same intensity-duration-frequency curve.
This study demonstrates how to determine critical runoff and infiltration events for a
single case study. However, the methodology presented in this study may be used for
determining critical rainfall information for other locations, soil textures, and intensity-
duration-frequency curves. This chapter provides a sound approach for determining
critical rainfalls within a given return period with the greatest potential for mobilization
of pathogens and the greatest risk to human health. This information is necessary for the
development of a comprehensive quantitative microbial risk assessment for exposure to
87
pathogens from land-applied soil amendments.
Chapter 3 provides useful information on microbial removal efficiencies of
subsurface media. The results can contribute to an assessment of the risk of water
contamination from land application of biosolids, and to determine safe setback distances
between disposal fields and receiving waters. Preferential flow pathways, as indicated by
variability in both observed and predicted flow velocity and dispersivity values, were
observed in this lysimeter study and may contribute to enhanced microbial transport.
Even under this extreme condition, the 2.4-meter sandy-loam soil is protective in terms of
reducing microbial indicators and removing viruses to below detection limits. It should
be noted that hydraulic parameter values, especially flow rate and dispersivity, are critical
to the performance of the model. The microbial parameters, such as retardation and decay
rate, are also specific to the study site conditions and flow patterns. P-22 performed well
as an indicator for the total mass of adenovirus transported through the porous media.
The performance of the microbial indicator is affected by retardations during transport,
and may miss the actual peak breakthrough time of the pathogen. The reliability of the
results from the subsurface fate and transport model can be improved by adjusting the
values of key parameters.
Chapter 4 describes a spreadsheet-based tool, the SMART Biosolids model,
which links quantitative microbial risk assessment with microbial fate and transport
modeling. The model combines spreadsheets with add-in visual basic macros in a rational
and supportable manner. Regulators and land application program managers may be able
to use the model to review different sites and determine which sites are most appropriate
for land application. Researchers may use the model to integrate information and identify
88
key gaps in knowledge warranting future research. Like many risk models, the SMART
Biosolids model requires numerous assumptions. Several key assumptions for the
groundwater pathway include the use of homogeneous media transport models and the
use of a fixed desorption fraction to describe the release of pathogens from the solid
phase to the aqueous phase. More details of model assumptions are provided in the
manual (Galada, Gurian et al. 2012a). In general the approach has been to be
conservative, that is, to err on the side of overestimating risk. Nevertheless, the impact
of different model structural assumptions is not always clear, and model risk estimates
may not be health protective in all cases. The default data in the spreadsheet model came
from various sources and may not be universally applicable.
Chapter 5 applied the SMART Biosolids model to assess microbial cumulative
risks over time. The risk from the groundwater pathway was compared to other exposure
pathways and the comparison of risks across pathogens was presented. It is found that
adenovirus presents the greatest risks across different pathways. Cryptosporidium and
Giardia lamblia produces the next highest risk levels except for air pathway. The risks
produced across pathways are ranked, in a descending order, as soil, surface water,
vegetables ingestion, air, and groundwater. The results from sensitivity analysis for risk
from adenovirus indicate that the uncertainties were contributed from different inputs
across different exposure pathways (we made the followed conclusion based on the
analysis for risk from adenovirus only). Microbial parameters, including initial
concentration in biosolids, release parameter, decay rates, and dose-response models, are
strongly correlated to the risk estimates. There is especially large uncertainty in risk
estimates for groundwater exposure pathway. Decay rate is especially important, which is
89
statistically significant for all pathways except for inhalation. Decay rate is also the only
input identified as significant for groundwater pathway. In order to reduce the big
uncertainties in risk estimation for groundwater exposure pathway, more field data is
required for microbial persistent in groundwater. Decay rate is also a key input for soil
pathway with a correlation of 0.98. However, in the SMART Biosolids model, the decay
in soil is assumed to be same as in water (Galada et al., 2012a; Galada et al., 2012b).
Although several studies observed a decay rate in soil close to the decay rate in
groundwater and suggested to use the same rates in soil and water (Lyon et al., 2001;
Torkzaban et al., 2006; Anders et al., 2009), more research needs to be done to compare
the decay rates in different environment. The hydraulic parameters, including hydraulic
conductivity, saturated water content, residual water content, and dispersion pattern, need
to match the specific environmental condition. It is found that coliphage and fecal
coliforms are very good indicators for all pathogen types, especially for inhalation and
surface water ingestion pathways. The risk estimated in this chapter is for residential
adults only. Health risks of other population can be examined using the same approach in
the future.
Quantitative uncertainty analysis is useful in setting regulatory limits, showing the
bounds of risk, showing the variability of risk across populations, exposure pathways,
and pathogens of concern, and finding the main contributor to those individuals facing
the highest risk. This new risk assessment compiles the most current pathogen content
data and exposure analysis. The assessment tool has the capability to archive the most up-
to-date knowledge and to be updated as additional information becomes available in the
future. If efforts can be made to reduce the uncertainties in risk estimations, this tool can
90
be used to improve original treatment technology requirements and setback regulations.
BIBLIOGRAPHY
Adams, B. L. and T. C. Bates (2003). Survival of Helicobacter pylori in a Natural
Freshwater Environment. Applied and Environmental Microbiology 69(12):
7462-7466.
Anders, R. and C. V. Chrysikopoulos (2009). Transport of Viruses Through Saturated
and Unsaturated Columns Packed with Sand. Transport in Porous Media, 76,
121–138.
Armstrong, T. W. and C. N. Haas (2007). Quantitative microbial risk assessment model
for Legionnaires' disease: assessment of human exposures for selected spa
outbreaks. Journal of Occupational and Environmental Hygiene 4(8): 634-646.
Arunraj, N. S. and J. Maiti (2009). A methodology for overall consequence modeling in
chemical industry. Journal of Hazardous Materials 169(1-3): 556-574.
Asano, T. (2007). Water Reuse: Issues, Technologies and Applications. New York, NY.,
McGraw-Hill Professional.
Azevedo, N. F. and C. Almeida (2008). Survival of Gastric and Enterohepatic
Helicobacter spp. in Water: Implications for Transmission. Applied and
Environmental Microbiology 74(6): 1805-1811.
Barth, G. R. and M. C. Hill (2005). Parameter and observation importance in modelling
virus transport in saturated porous media—investigations in a homogenous
system. Journal of Contaminant Hydrology, 80(3-4), 107–129.
doi:10.1016/j.jconhyd.2005.06.012
Bitton, B., O. C. Pancorbo, and S. R. Farrah (1984). Virus Transport and Survival After
Land Application of Sewage Sludge. Applied and Environmental Microbiology
47(5): 905-909.
Bouwknegt, M., S. A. Rutjes, and A. M. de Roda Husman (2009). Hepatitis E virus risk
profile: Identifying potential animal, food and water sources for human
infections. . RIVM: Bilthoven, Netherlands., National Institute for Public Health
and the Environment: 1-60.
Bradford, S. A., J. Simunek, M. Bettahar, M. T. van Genuchten, and S. R. Yates (2003).
Modeling colloid attachment, straining, and exclusion in saturated porous media.
Environmental Science & Technology 37(10): 2242-2250.
Brooks, J. P. and C. P. Gerba (2004). Biological Aerosol Emission Fate and Transport
from Municipal and Animal Wastes. J. Residuals Sci. Technol. 1: 16-28.
92
Brooks, J. P., B. D. Tanner, C. P. Gerba, C. N. Haas, and I. L. Pepper (2005a). Estimation
of Bioaerosol risk of Infection to Residents Adjacent to a Land Applied Biosolids
Site Using an Empirically Derived Transport Model. Journal of Applied
Microbiology, 98:397-405.
Brooks, J. P., B. D. Tanner, C. Gerba, C. Haas, and I. Pepper (2005b). Estimation of
bioaerosol risk of infection to residents adjacent to a land applied biosolids site
using an empirically derived transport model. Journal of Applied Microbiology
98(2): 397-405.
Brooks, J. R., B. D. Tanner, K. L. Josephson, C. P. Gerba, C. N. Haas, and I. L. Pepper
(2005c). A national study on the residential impact of biological aerosols from the
land application of biosolids. The Journal of Applied Microbiology 99: 310-322.
Butler, R. G., G. T Orlob. and P. H. McGaughey (1954). Underground movement of
bacterial and chemical pollutants. J. Amer. Water Works Assoc. 46: 97-111.
Carrasco, I. J. and S. Y. Chang (2005). Random Monte Carlo simulation analysis and risk
assessment for ammonia concentrations in wastewater effluent disposal.
Stochastic Environmental Research and Risk Assessment 19(2): 134-145.
Chacin-Bonilla, L. (2010). Epidemology of Cyclospora: A review focusing in endemic
areas. Acta Tropica 115: 181-193.
Chauret, C., S., S. Springthorpe, and S. Sattar (1999). Fate of Cryptosporidium oocysts,
Giardia cysts, and microbial indicators during wastewater treatment and anaerobic
sludge digestion. Canadian Journal of Microbiology 45(3): 257-262.
Chetochine, A. S., M. L. Brusseau, C. P. Gerba, and I. L. Pepper (2006). Leaching of
phage from Class B Biosolids and potential transport through soil. Applied and
Environmental Microbiology 72(1): 665-671.
Choun, Y. S. and A. S. Elnashai (2010). A simplified framework for probabilistic
earthquake loss estimation. Probabilistic Engineering Mechanics 25(4): 355-364.
Cook, K. L. and C. H. Bolster (2007). Survival of Campylobacter jejuni and Escherichia
coli in groundwater during prolonged starvation at low temperatures. Journal of
Applied Microbiology 103(3): 573-583.
Corapcioglu, M. Y. and A. Haridas (1984). Transport and fate of microorganisms in
porous media: A theoretical investigation. J. Hydrology (72);149-169.
Cullen, A. C., and H. C. Frey (1999). Probabilistic Techniques in Exposure Assessment.
Plenum Press: New York.
93
Curriero, F. C., J. A. Patz, J. B. Rose, and S. Lele (2001). The Association Between
Extreme Precipitation and Waterborne Disease Outbreaks in the United States,
1948–1994. American Journal of Public Health 91(8): 1194–1199.
Dixon, K. L., P. L. Lee, and G. P. Flach (2008). A graded approach to flow and transport
modeling to support decommissioning activities at the Savannah River Site.
Health Physics 94(5): S56-S61.
Donald, M., K. Mengersen, S. Toze, J. P. Sidhu, and A. Cook (2011). Incorporating
parameter uncertainty into Quantitative Microbial Risk Assessment (QMRA).
Journal of Water and Health 9(1): 10-26.
Dowd, S. E., C. P. Gerba, I. L. Pepper, S. D. Pillai (2000). Bioaerosol transport modeling
and risk assessment in relation to biosolid placement. Journal of Environmental
Quality 29: 343-348.
Doyle, M. P. (2001). Predictive Modeling and Risk Assessment. Food Microbiology:
Fundamentals and Frontiers. Washington, D.C., ASM Press.
Drayna, P., McLellan, S. L., Simpson, P., Li, S.-H., and Gorelick, M. H. (2010).
Association between Rainfall and Pediatric Emergency Department Visits for
Acute Gastrointestinal Illness. Environmental Health Perspectives, 118(10),
1439–1443.
Dubey, J. P. (1998). Toxoplasma gondii Oocyst Survival under Defined Temperatures.
The Journal of Parasitology 84(4): 862-865.
Eisenberg, J. N. (2006). Application of a Dynamic Model to Assess Microbial Health
Risks Associated with Beneficial Uses of Biosolids. Water Environment Research
Foundation, 98-REM-1a: 1-124.
Eisenberg, J. N., K. Moore, J. A. Soller, D. Eisenberg, and J. M. Colford Jr. (2008).
Microbial Risk Assessment Framework for Exposure to Amended Sludge Projects.
Environmental Health Perspectives 116(6): 1-7.
Enriquez, C. E., C. J. Hurst, and C. P. Gerba (1995). Survival of The Enteric Adenovirus-
40 and Adenovirus-41 In Tap, Sea, and Waste-Water. Water Research 29(11):
2548-2553.
Environmental Protection Department (2005). Chapter 406. Natural and Historic
Resources Protection; Article 12. Wastewater Treatment Facilities. Alachua
County, Florida Retrieved from
http://www.alachuacounty.us/Depts/EPD/WaterResources/Documents/Article%2
012%20wastewater%20treatment%20facilities.pdf.
Erickson, M. C. and Y. R. Ortega (2006). Inactivation of protozoan parasites in food,
water, and environmental systems. Journal of Food Protection 69(11): 2786-2808.
94
Espinosa, A. C. and M. Mazari-Hiriart (2008). Infectivity and genome persistence of
rotavirus and astrovirus in groundwater and surface water. Water Research 42:
2618-2628.
Evanylo, G. K. (2009). Agricultural Land Application of Biosolids in Virginia:
Regulations, Retreived from http://pubs.ext.vt.edu/452/452-302/452-302.html.
Faulkner, B. R., W. G. Lyon, and F. A. Khan (2002). Predicting attenuation of viruses
during percolation in soils. Cincinnati, OH 45268, National Risk Management
Research Laboratory, Office of Research and Development, U.S. Environmental
Protection Agency.
Feachem, R. G., D. J. Bradley, H. Garelick, and D. D. Mara (1983). Sanitation and
disease: Health aspects of excreta and wastewater management. Chichester: John
Wiley & Sons.
Filip, Z. and D. Kaddu-Mulindwa (1988). Survival of some pathogenic and facultative
pathogenic bacteria in groundwater. Water Science and Technology 20(3): 227-
231.
Foppen, J. W., M. Van Herwerden and J. Schijven (2007). Measuring and modelling
straining of Escherichia coli in saturated porous media. Journal of Contaminant
Hydrology 93(1-4): 236-254.
Fout, G. S., B. C. Martinson, M. W. N. Moyer, and D. R. Dahling (2003). A Multiplex
Reverse Transcription-PCR Method for Detection of Human Enteric Viruses in
Groundwater. Applied and Environmental Microbiology 69(6): 3158–3164.
Frey, H. C., and S. R. Patil (2002). Identification and Review of Sensitivity Analysis
Methods. Risk Analysis 22(3): 1–26.
Galada, H., P. L. Gurian, A. Joe, A. kumar, B. Olson, M. S. Olson, E. Richter, J. Teng,
and H. Zhang (2012a). Site Specific Risk Assessment Tools for Land Applied
Biosolids. Water Environment Research Foundation. SRSK3R08.
Galada, H., P. L. Gurian, T. Hong, A. Johnson, A. Kumar, M. S. Olson, E. Richter, N.
Rostad, J. Teng, K. Wong, I. Xagoraraki, and E. Casman (2012b). Calibrating and
Applying the Spreadsheet Model of Risk: Tool for Biosolids to Fault Scenarios,
Water Environment Research Foundation. SRSK3R08.
Gale, P. (2005). Land application of treated sewage sludge: quantifying pathogen risks
from consumption of crops. Journal of Applied Microbiology 98(2): 380-396.
Gerba, C. P., I. L. Pepper, and L. F. Whitehead (2002). A risk assessment of emerging
pathogens of concern in the land application of biosolids. Water Science and
Technology 46(10): 225-230.
95
Gerba, C. P. (2010) Protozoan pathogens in air. Personal email and phone
communication.
Ginn, T. R., B. D. Wood, K. E. Nelson, and T. D. Scheibe (2002). Processes in microbial
transport in the natural subsurface. Advances in Water Resources 25(8-12): 1017–
1042.
Griffiths, H. J. (1978). A handbook of veterinary parasitology of domestic animals of
North America, University of Minnesota, Minneapolis, MN.
Guber, A. K., Y. A. Pachepsky, A. M. Yakirevich, D. R. Shelton, A. M. Sadeghi, D. C.
Goodrich, and C. L. Unkrich (2011). Uncertainty in modelling of faecal coliform
overland transport associated with manure application in Maryland. Hydrological
Processes 25(15): 2393–2404.
Guzman, C., J. Jofre, M. Montemayor, and F. Lucena (2007). Occurrence and levels of
indicators and selected pathogens in different sludges and biosolids. Journal of
Applied Microbiology 103: 2420-2429.
Haas, C. N., J. B. Rose, and C. P. Gerba (1999). Quantitative Microbial Risk Assessment,
John Wiley & Sons, Inc. New York, NY.
Hagedorn, C. (1983). Bacterial pathogens in groundwater, in microbial health
considerations of soil disposal of domestic wastewaters. EPA-600/9-83-017:153-
171. Municipal Env. Res. Lab. Cincinnati, Ohio.
Hammonds, J., and F. Hoffman (1994). An introductory guide to uncertainty analysis in
environmental and health risk assessment. US DOE.
Harrison, E. Z., S. R. Oakes, and M. Hysell (2006). Organic chemicals in sewage sludges.
Science of the Total Environment 367(2-3): 481-497.
Helton, J. C. (1993). Uncertainty and sensitivity analysis techniques for use in
performance assessment for radioactive waste disposal. Reliability Engineering
and System Safety 42: 327–367.
Helton, J. C., J. D. Johnson, C. J. Sallaberry, and C. B. Storlie (2006). Survey of
sampling-based methods for uncertainty and sensitivity analysis. Reliability
Engineering and System Safety 91(10-11): 1175–1209.
Henis, Y. (1987). Survival and dormancy of microorganisms. New York, Wiley-
Interscience.
Hutter, S. E. and U. Kihm (2010). Assessment of exposure to bovine spongiform
encephalopathy in a hypothetical country. Revue Scientifique Et Technique-
Office International Des Epizooties 29(3): 459-471.
96
Jackson, G. T., J. W. Bier, and R. A. Rude (1977). Recycling of refuse into the food
chain: The parasite problem. Risk assessment and health effects of land
application of municipal wastewater and sludges. B. P. Sagik and C. A. Sorber,
Center for Appl Res and Tech, University of Texas, San Antonio, TX.
John, D. E. and J. B. Rose (2005). Review of factors affecting microbial survival in
groundwater. Environmental Science & Technology 39(19): 7345-7356.
Juneja, V. K. and H. Marks (2001). Discussion of Non-linear survival curves. Presented
at Society of Risk Analysis Workshop, Decision Support Tools for Microbial Risk
Assessment, Annual Meeting, Seattle, WA.
Juneja, V. K. and H. M. Marks (2003). Mathematical description of non-linear survival
curves of Listeria monocytogenes as determined in a beef gravy model system at
57.5 to 65 C. Innovative Food Science & Emerging Technologies 4(3): 307-317.
Kamra, S. K., B. Lennartz, and M. T. van Genuchten (2001). Evaluating non-equilibrium
solute transport in small soil columns. Journal of Contaminant Hydrology 48(3-4):
189-212.
Keswick, B. H., C. P. Gerba, and S. L. Secor (1982). Survival of Enteric Viruses and
Indicator Bacteria In Groundwater. Journal of Environmental Science and Health
Part a-Environmental Science and Engineering & Toxic and Hazardous Substance
Control 17(6): 903-912.
Kim, J. and X. Jiang (2010). The growth potential of Escherichia coli O157:H7,
Salmonella spp. and Listeria monocytogenes in dairy manure-based compost in a
greenhouse setting under different seasons. Journal of Applied Microbiology
109(6): 2095-2104.
Kingsley, D. H., D. R. Hollinian, and K. R. Calci (2007). Inactivation of a norovirus by
high-pressure processing. Applied and Environmental Microbiology 73(2): 581-
585.
Knightes, C. D. (2008). Development and test application of a screening-level mercury
fate model and tool for evaluating wildlife exposure risk for surface waters with
mercury-contaminated sediments (SERAFM). Environmental Modelling &
Software 23(4): 495-510.
Kopylev, L., C. Chen, and P. White (2007). Towards quantitative uncertainty assessment
for cancer risks: Central estimates and probability distributions of risk in dose-
response modeling. Regulatory Toxicology and Pharmacology 49(3): 203-207.
Koudela, B., S. Kucerova, and T. Hudeovic (1999). Effect of low and high temperatures
on infectivity of Encephalitozoon cuniculi spores suspended in water. Folia
Parasitologica 46(3): 171-174.
97
Lathem, W. W., S. D. Crosby, V. L. Miller, and W. E. Goldman (2005). Progression of
primary pneumonic plague: A mouse model of infection, pathology, and bacterial
transcriptional activity. Proceedings of the National Academy of Sciences of the
United States of America 102(49): 17786-17791.
Levy, J., K. Sun, R. H. Findlay, F. T. Farruggia, J. Porter, K. L. Mumy, and J. Tomaras
(2007). Transport of Escherichia coli bacteria through laboratory columns of
glacial-outwash sediments: Estimating model parameter values based on sediment
characteristics. Journal of Contaminant Hydrology 89(1-2): 71–106.
Li, B. L., C. Loehle, and D. Malon (1996). Microbial transport through heterogeneous
porous media: Random walk, fractal, and percolation approaches. Ecological
Modelling 85(2-3): 285–302.
Lindqvist, R. and A. Westoo (2000). Quantitative risk assessment for Listeria
monocytogenes in smoked or gravad salmon and rainbow trout in Sweden.
International Journal of Food Microbiology 58(3): 181-196.
Lyon, W. G. and B. R. Faulkner (2001). Identification of parameters for predicting fate
and transport of viruses in soils and sediments. Ada, Oklahoma 74820, ManTech
Environmental Research Services Corporation.
Macler, B. A., and J. C. Merkle. (2000). Current knowledge on groundwater microbial
pathogens and their control. Hydrogeology Journal 8: 29–40.
Mantovi, P., G. Baldoni, and G. Toderi (2005). Reuse of liquid, dewatered, and
composted sewage sludge on agricultural land: effects of long-term application on
soil and crop. Water Research 39(2-3): 289-296.
Mara, D. and A. Sleigh (2010). Estimation of Ascaris infection risks in children under 15
from the consumption of wastewater-irrigated carrots. Journal of Water and
Health 08(1): 35-38.
McFeters, G. A. and G. K. Bissonnette (1974). Comparative Survival of Indicator
Bacteria and Enteric Pathogens in Well Water. Applied Microbiology 27(5): 823-
829.
Medema, G. J., F. M. Schets, P. F. M. Teunis, and A. H. Havelaar (1998). Sedimentation
of Free and Attached Cryptosporidium Oocysts and Giardia Cysts in Water.
Applied and Environmental Microbiology 64(11): 4460-4466.
Metzger, J. N., R. A. Fjeld, J. S. Hammonds, and F. O. Hoffman (1998). Evaluation of
software for propagating uncertainty through risk assessment models. Ecol. Risk
Assess 4(2): 263–290.
Morgan, M. (1992). Uncertainty: a guide to dealing with uncertainty in quantitative risk
and policy analysis. Cambridge: Cambridge University Press.
98
Nasser, A. M., and S. D. Oman (1999). Quantitative assessment of the inactivation of
pathogenic and indicator viruses in natural water sources. Water Research 33(7):
1748–1752.
National Research Council (2002). Biosolids Applied to Land-Advancing Standards and
Practices. Washington, D.C., The National Academies Press.
Nauta, M. J. (2005). Microbiological risk assessment models for partitioning and mixing
during food handling. International Journal of Food Microbiology 100(1-3): 311-
322.
Ngazoa, E. S., I. Fliss, J. Jean (2008). Quantitative study of persistence of human
norovirus genome in water using TaqMan real-time RT-PCR. Journal of Applied
Microbiology 104(3): 707-715.
Oscar, T. P. (2002). Development and validation of a tertiary simulation model for
predicting the potential growth of Salmonella typhimurium on cooked chicken.
International Journal of Food Microbiology 76(3): 177-190.
Oscar, T. P. (2004). A quantitative risk assessment model for Salmonella and whole
chickens. International Journal of Food Microbiology 93(2): 231-247.
Overcash, M., R. C. Sims, J. L. Sims (2005). Beneficial reuse and sustainability: The fate
of organic compounds in land-applied waste. Journal of Environmental Quality
34(1): 29-41.
Pang, L. (2009). Microbial Removal Rates in Subsurface Media Estimated From
Published Studies of Field Experiments and Large Intact Soil Cores. Journal of
Environmental Quality 38(4): 1531–1559.
Park, H. S. and C. San Juan (2000). A method for assessing leaching potential for
petroleum hydrocarbons release sites: Multiphase and multisubstance equilibrium
partitioning. Soil & Sediment Contamination 9(6): 611-632.
Pepper, L. L., J. P. Brooks, and C. P. Gerba (2006). Pathogens in biosolids. Advances in
Agronomy, 90: 1-41.
Pepper, I. L., J. P. Brooks, R. G. Sinclair, P. L. Gurian, and C. P. Gerba (2010).
Pathogens and Indicators in United States Class B Biosolids: National and
Historic Distributions. Journal of Environmental Quality 39: 1-6.
Pitblado, R. (1994). Quality and Offshore Quantitative Risk Assessment. Journal of Loss
Prevention in the Process Industries 7(4): 360-369.
Ramaiah, N., V. Kolhe, A. Sadhasivan (2004). Abundance of pollution indicator and
pathogenic bacteria in Mumbai waters. Current Science 87(4): 435-439.
99
Regli, S., J. B. Rose, C. N. Haas, and C. P. Gerba (1991). Modeling the risk from giardia
and viruses in drinking-water. Journal American Water Works Association 83(11):
76-84.
Risebro, H. L., P. R. Hunter, M. F. Doria, Y. Andersson, G. Medema, K. Osborn, and O.
Schlosser (2007). Fault tree analysis of the causes of waterborne outbreaks.
Journal of Water and Health, 5(Suppl 1), S1.
Rose, J. B., S. Daeschner, and D. R. Easterling (2000). Climate and waterborne disease
outbreaks. Journal American Water Works Association 92(9): 77-87.
Ross, T. and J. Sumner (2002). A simple, spreadsheet-based, food safety risk assessment
tool. International Journal of Food Microbiology 77(1-2): 39-53.
Rucker, G. G. (2007). Vadose Zone Contaminant Migration Model - Muti-Layered
Software© (VZCOMML) Calculation Note and Documentation. Washington
Savannah River Company. Aiken, SC 29808.
Schijven, J. F., S. M. Hassanizadeh, R. H. A. M. de Bruin (2002). Two-site kinetic
modeling of bacteriophages transport through columns of saturated dune sand.
Journal of Contaminant Hydrology 57(3-4): 259–279.
Schijven, J. F. (2003). Bacteriophages and clostridium spores as indicator organisms for
removal of pathogens by passage through saturated dune sand. Water Research 37:
2186–2194.
Seiler, F. A., and J. L. Alvarez (1996). On the selection of distributions for stochastic
variables. Risk Analysis 16(1):5-18.
Sidhu, J. P. S. and S. G. Toze (2009). Human pathogens and their indicators in biosolids:
A literature review. Environment International,35(1): 187-201.
Singh, R. P. and M. Agrawal (2008). Potential benefits and risks of land application of
sewage sludge. Waste Management 28(2): 347-358.
Sinton, L. W., R. K. Finlay, L. Pang, and D. M. Scott (1997). Transport of bacteria and
bacteriophages in irrigated effluent into and through an alluvial gravel aquifer.
Water, Air, and Soil Pollution 98(1-2): 17–42.
Sinton, L. W., M. L. Mackenzie, N. Karki, R. R. Braithwaite, C. H. Hall, and M. J.
Flintoft (2010). Transport of Escherichia coli and F-RNA bacteriophages in a 5m
column of saturated pea gravel. Journal of Contaminant Hydrology 117(1-4): 71–
81.
Smith, M. A., K. Takeuchi, G. Anderson (2008). Dose-response model for Listeria
monocytogenes-induced stillbirths in nonhuman primates. Infection and Immunity
76(2): 726-731.
100
Soller, J. A. (2004). Evaluation of Microbial Risk Assessment Techniques and
Applications. Alexandria, VA, Water Environment Research Foundation. Report
00-PUM-3.
Sumner, J. and T. Ross (2002). A semi-quantitative seafood safety risk assessment.
International Journal of Food Microbiology 77(1-2): 55-59.
Tanner, B. D., J. P. Brooks, C. P. Gerba, C. N. Haas, K. L. Josephson, and I. L. Pepper
(2008). Estimated occupational risk from bioaerosols generated during land
application of class B biosolids. J. Environ. Qual. 37: 2311-2321.
Teng, J. (2010). Determination of Critical Rainfall Events for Quantitative Microbial
Risk Assessment of Biosolids-Associated Pathogens. Residuals and Biosolids
2010, 1–16.
Teng, J., A. Kumar, H. Zhang, M. S. Olson and P. L. Gurian (2012a). Determination of
Critical Rainfall Events for Quantitative Microbial Risk Assessment of Land-
Applied Soil Amendments. Journal of Hydrologic Engineering 17(3): 437-444.
Teng, J., M. S. Olson, P. L. Gurian, T. Harrigan, A. Kumar, K. Wong, and I. Xagoraraki
(2012b). Validation of Microbial Transport in Porous Media with Field
Monitoring Data. Environmental Science & Technology (in preparation).
Teng, J., A. Kumar, P. L. Gurian, and M. S. Olson (2012c). A Spreadsheet-Based Site
Specific Risk Assessment Tool for Land-Applied Biosolids. Environmental
Modeling & Software (second round review).
Teunis, P. F. M., O. G. van der Heijden, J. W. B. van der Giessen, and A. H. Havelaar
(1996). The dose-response relation in human volunteers for gastro-intestinal
pathogens. Rijksinstituut voor Volksgezondheid en Milieu RIVM: 1-97.
Teunis, P. F. M., and F. Kasuga (2010). Dose-response modeling of Salmonella using
outbreak data. International Journal of Food Microbiology 144(2): 243-249.
Torkzaban, S., S. M. Hassanizadeh, J. F. Schijven, H. A. M. de Bruin, and A. M. de Roda
Husman (2006). Virus Transport in Saturated and Unsaturated Sand Columns.
Vadose Zone Journal 5: 877–885.
Trefry, C. M., D. W. Watkins Jr,, and D. Johnson (2005). Regional Rainfall Frequency
Analysis for the State of Michigan. Journal of Hydrologic Engineering 10(6):
437–449.
Tufenkji, N. (2007). Modeling microbial transport in porous media: Traditional
approaches and recent developments. Advances in Water Resources 30(6-7):
1455-1469.
101
Turner, J. E., H. A. Wright, and R. N. Hamm (1985). A Monte Carlo primer for health
physicists. Health physics 48(6): 717-733.
Tye, B. K., R. K. Chan, and D. Botstein (1974). Packaging of an oversize transducing
genome by Salmonella phage P22. Journal of Molecular Biology 85(4): 485-490.
U.S. EPA (1994). Land Application of Sewage Sludge: A Guide for Land Appliers on the
Requirements of the Federal Standard for the Use or Disposal of Sewage Sludge.
Washington, D.C. 40 CFR Part 503; EPA/831-B-93-002b.
U.S. EPA (2000). A Guide to Field Storage of Biosolids and the Organic By-Products
Used in Agriculture and for Soil Resource Management. EPA/832-B-00-007,
Washington, DC.
U.S.EPA (1989). Pathogen Risk Assessment for Land Application of Municipal Sludge.
Volume I: Methodology and computer model. W. U.S. Environmental Protection
Agency, D.C.
U.S.EPA (1992). Technical Support Document for Reduction of Pathogens and Vector
Attraction in Sewage Sludge. U. S. E. P. A. Office of Health and Environmental
Assessment, Washington, D.C.
U.S.EPA (1995). A Guide to the Biosolids Risk Assessment for the EPA Part 503 Rule.
U.S. W. U.S. Environmental Protection Agency, D.C.
Vandeven, M., T. Whitaker, and A. Slate (2002). Statistical approach for risk assessment
of aflatoxin sampling plan used by manufacturers for raw shelled peanuts. Journal
of Aoac International 85(4): 925-932.
Viau, E. and J. Peccia (2009). Survey of Wastewater Indicators and Human Pathogen
Genomes in Biosolids Produced by Class A and Class B Stabilization Treatments.
Applied and Environmental Microbiology 75(1): 164-174.
Viau, E., K. Bibby, T. Paex-Rubio, and J. Peccia (2011). Toward a Consensus View on
the Infectious Risks Associated with Land Application of Sewage Sludge.
Environmental Science & Technology 45(13): 5459-5469.
Virginia Department of Environmental Quality (2010). Guidelines fro the Land
Application and Surface Disposal of Biosolids, Retrieved from
http://www.deq.state.va.us/export/sites/default/info/documents/biosolids/TN_BS_
GuidelinesPRINT120600_Rev051101.pdf.
Vose, D. J. (1998). The application of quantitative risk assessment to microbial food
safety. Journal of Food Protection 61(5): 640-648.
102
Waddill, D. W., and M. A. Widdowson (1998). Three-dimensional model for subsurface
transport and biodegradation. Journal of Environmental Engineering - ASCE
124(4): 336–344.
Westrell, T. (2004). QMRA (quantitative microbial risk assessment) and HACCP (hazard
analysis and critical control points) for management of pathogens in wastewater
and sewage sludge treatment and reuse. Water Science and Technology 50(2): 23-
30.
Wong, K., B. Onan, and I. Xagoraraki (2010). Quantification of enteric viruses,
indicators and salmonella in Class B anaerobic digested biosolids by culture and
molecular methods. Applied and Environmental Microbiology 76(19): 6441-6448.
Xagoraraki, I., A. Kumar, and K. Wong (2010). Michigan State University Biosolids-
team.
103
APPENDIX
Table A-1 Compilation of occurrence by pathogen (items in red are data gaps) (Chapter 1)
Pathogen
Distribution type
("1" for
"Normal", "2" for
"Uniform", "0"
for not
applicable)
Mean Standard
deviation Minimum Maximum Reference
Cryptosporidium1 2.00E+00 2.80E+01 1.90E+01 1.30E+01 6.40E+01 Guzman et al. (2007)
Cyclosporidia 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Entamoeba
histolytica 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Giardia lamblia2 2.00E+00 1.28E+01 2.00E-19 2.50E-13 2.82E+01 Chauret et al. (1999)
Microsporidia 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Campylobacter
jejuni4
2.00E+00 <1.00E+00 2.00E-01 1.00E-20 1.00E+01 Pepper et al. (2010)
Clostridium spp. 2.00E+00 4.16E+07 1.86E+08 3.98E+04 8.53E+08 Pepper et al. (2010)
E.coli O1573 2.00E+00 <1.00E+00 2.00E-01 1.00E-20 1.00E+01 Pepper et al. (2010)
Helicobacter 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Listeria 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Salmonella
spp.5,6
2.00E+00 8.10E-01 2.60E+00 2.50E-21 3.35E-01 Pepper et al. (2010)
Shigella spp.5 2.00E+00 4.49E+00 5.37E+01 1.00E-20 2.00E+00 Pepper et al. (2010)
104
Pathogen
Distribution type
("1" for
"Normal", "2" for
"Uniform", "0"
for not
applicable)
Mean Standard
deviation Minimum Maximum Reference
Vibrio cholera 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Yersinia spp. 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Adenovirus 2.00E+00 1.76E+01 1.33E+01 3.70E+00 2.26E+01 Pepper et al. (2010)
Ascaris4,6
2.00E+00 <2.50E-01 5.00E-02 2.50E-21 2.50E+00 Pepper et al. (2010)
Coliphage5
(Somatic) 2.00E+00 8.40E+08 3.38E+12 1.00E-20 1.92E+07 Pepper et al. (2010)
Enteroviruses5,6
2.00E+00 1.05E-01 2.00E-01 1.38E-02 8.00E-01 Pepper et al. (2010)
Hepatitis A virus 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Hepatitis E 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Astrovirus 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Legionella 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Norovirus 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Rotavirus 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Toxoplasma 1.00E+00 1.00E-18 2.00E-19 1.00E-20 1.00E-17
Fecal coliforms 2.00E+00 1.27E+07 3.62E+07 5.17E+01 1.58E+08 Pepper et al. (2010)
E-coli 2.00E+00 3.16E+03 2.00E+01 6.05E+00 1.12E+06 Wong et al. (2010)
Enterococci7 2.00E+00 1.27E+04 1.26E+01 1.00E-20 3.15E+05 Pepper et al. (2010)
1Densities per 10 g of DM; All initial values have been divided by 10.
2Cake Product -Organisms per 100g of wet sludge; Counts corrected to account for dewatering; All initial value have been divided by
100.
105
3This is the detection limit
4All samples were below detection (<1).
5Some samples were below detection (<1); MLE values used.
6Organism per 4g; Minimum value was below detection (<1); All initial values have been divided by 4.
7Utilized values of Fecal streptococcus; Minimum value was below detection (<1).
106
Table A-2 Microbial decay in ground water (units in 1/ hour) (Chapter 1)
Pathogen Mean Standard
deviation Minimum Maximum Reference Remarks
Cryptosporidium 1.58E-03 Medema, 1998
Survival in water from sedimentation
experiments at 23 degree C (approximated
from Figure 1B in the reference paper)
Cyclosporidia 0.00E+00 Erickson, 2006 0% inactivation in 7 days at 20-25 degree C
Entamoeba
histolytica 4.80E-03 3.20E-03 6.40E-03 Feachem, 1983
Survival for 15 to 30 days in fresh water at
20 to 30 degree C (assumed to be 90%
inactivation)
Giardia lamblia 3.80E-03 Medema, 1998
Survival in water from sedimentation
experiments at 23 degree C (approximated
from Figure 1B in the reference paper)
Microsporidia 1.31E-04 Koudela, 1999
Survival of Encephalitozoon cuniculi
Levaditi in distilled water at 4 degree C for
2 years
Campylobacter
jejuni
3.36E-02 2.59E-03 Cook, 2007 Decay for LR underground river (from
Table 2 in the reference paper)
2.3E-02 Azevedo, 2008
Survival in water (assumed to be 90%
inactivation) at 25 C in the absence of light
(approximated from Figure 1 in the
reference paper)
Clostridium spp. 2.50E-04 Filip, 1988
Survival of clostridium perfringens in
groundwater (approximated from Figure 1
in the reference paper)
E.coli O157 2.88E-02 3.84E-02 9.59E-04 1.44E-01 John, 2005
Assumed to be same as coliform bacteria at
temp from 3 to 37 degree C (from Table 2
in the reference paper)
107
Pathogen Mean Standard
deviation Minimum Maximum Reference Remarks
Helicobacter
3.1E-02 1.4E-02 4.8E-02 Azevedo, 2008
Survival in water (assumed to be 90% inactivation)
at 25 C in the absence of light. Minimum from H.
pylori 968, and maximum from H. mustelae and H.
muridarum (approximated from Figure 1 in the
reference paper)
2.45E-01 3E-02 4.6E-01 Adams, 2003
Survival time under temperature from 16 C to 23
C in natural fresh water environment (assumed to
be 90% inactivation) (approximated from Figure 3
in the reference paper)
Listeria 3.43E-03 Kim, 2010 Survival of 28 days in manure-based compost
(assumed to be 90% inactivation)
Salmonella
spp. 9.59E-03 1.91E-02 2.88E-03 5.75E-02 John, 2005
At temperature range from 10 to 22 degree C (from
Table 2 in the reference paper)
Shigella spp. 4.40E-03 Henis, 1987 Survival time is 22 days in wells (assumed to be
90% inactivation)
Vibrio
cholera 2.27E-03 Ramaiah, 2004
Survival in natural, filtered seawater (from
starvation duration of 75 days in Table 3 of the
reference paper)
Yersinia spp. 5.00E-04 Filip, 1988 Yersinia enterocolitica in groundwater
(Approximated from Figure 1 in the reference paper)
Adenovirus 1.75E-03 Enriquez, 1995
Persistence in secondary sewage effluent at 15
degree C; Average of adeno 40 and 41 (from Table 3
in the reference paper)
Ascaris 7.67E-05 6.57E-05 8.76E-05
Jackson, 1977
& Griffiths,
1978
Survival of 3 to 4 years in soil (assumed to be 90%
inactivation)
Coliphage 2.88E-03 2.88E-03 2.30E-10 9.58E-03 John, 2005 At temperature from 0 to 10 degree C (from Table 3
in the reference paper)
108
Pathogen Mean Standard
deviation Minimum Maximum Reference Remarks
Enteroviruses 5.00E-03
Lyon &
Chattopadhyay
, 2001
Hepatitis A
virus 1.92E-03 3.84E-03 2.30E-10 7.66E-03 John, 2005
At temperature from 0 to 10 degree C (from Table 3
in the reference paper)
Hepatitis E
virus 1.92E-03 3.84E-03 2.30E-10 7.66E-03 John, 2005 Assumed to be same and HAV-A
Astrovirus 2.4E-03 Espinosa, 2008 Survival in groundwater (approximated from Figure
2 in the reference paper)
Norovirus 3E-04 1.6E-04 4.39E-04 Ngazoa, 2007
Survival in river at 4 degree C (from viral reduction
at 20 days and 30 days in Table 2 of the reference
paper)
Rotavirus 1.14E-03 Espinosa, 2008 Survival in groundwater (approximated from Figure
1 in the reference paper)
Toxoplasma 8.75E-03 1.5E-03 1.6E-02 Dubey, 1998
Survival at temperature from 35 to 55 degree C
(assumed to be 90% inactivation) (from Table 1 in
the reference paper)
Fecal
coliforms 1.88E-02
McFeters,
1974
E.coli 2.88E-02 3.84E-02 9.59E-04 1.44E-01 John, 2005 Table 2. Assumed to be same as coliform bacteria,
temp range is 3-37 degree C
Pathogen Mean Standard
deviation Minimum Maximum Reference Remarks
Enterococci 2.88E-02 2.88E-02 9.58E-04 7.67E-02 John, 2005 Table 2. Assumed to be same as coliform bacteria,
temp range is 3-22 degree C
109
Table A-3 Microbial partitioning values (Chapter 1)
Microorganisms Mean Minimum Maximum Reference Remarks
Adenovirus 2.5% 2% 4% Xagoraraki, 2010 Percentage of virus desorbed from soil (with 8%
organic matter) by the first extraction.
<1% Xagoraraki, 2010 Percentage of virus desorbed from soil (with 2%
organic matter) by the first extraction.
Coliphage 7.4% Chetochine, 2006 Recovery from column transport studies with 7%
biosolids.
4.3% 3.3% 5.3% Chetochine, 2006 Recovery from column transport studies with 2%
biosolids.
Poliovirus and
echovirus 0% Bitton, 1984
Percentage of virus in soil leachates collected after
natural rainfall
Salmonella
enterica ssp.
Enterica serovar
Thphimurium-lux
30% 13.88% 52.26% Horswell, 2008 Percentage of salmonella in leachate from sewage
sludge (200 kg N ha-1
) (study for New Zealand)
Adenovirus 0.0774% 0.0432% 0.1214% Horswell, 2008 Percentage of adenovirus in leachate from sewage
sludge (200 kg N ha-1
) (study for New Zealand)
110
Table A-4 Dose-response models for different biosolids-associated bacteria
Pathogen
D-R Type
Mean UF UB Reference (“1” for Exp, “2”
for Beta”, “0”
for Not Available)
Cryptosporidium 1.00E+00 4.19E-03 1.81E+00 95th
: 7.57E-03 Haas et al. (1999)
Cyclosporidia 1.00E+00 2.19E-02 3.16E+00 90th
: 6.93E-02 Chacin-Bonilla (2010)
Entamoeba
histolytica 1.00E+00 4.90E-02 1.41E+00 90
th: 6.93E-02 Asano et al. (2007)
Giardia lamblia 1.00E+00 2.00E-02 2.84E-02 95th
: 5.66E-02 Teunis et al. (1996); Haas et al (1999)
Microsporidia 0.00E+00 0.00E+00
Campylobacter
jejuni 2.00E+00 1.91E-02 2.32E+00 95
th: 4.43E-02 Haas et al. (1999); Teunis et al. (1996)
Clostridium spp. 1.00E+00 0 0 0
E.coli O157 2.00E+00 1.00E-07 1.00E+01 90th
: 1.00E-06 Haas et al. (1999)
Helicobacter 1.00E+00 Clapham et al. (2004)
Listeria 1.00E+00 1.76E-08 1.08E+01 95th
: 1.91E-07 Smith et al. (2008)
Salmonella spp. 2.00E+00 2.71E-06 1.14E+02 90th
: 3.09E-04 Soller et al. (2004)
Shigella spp. 2.00E+00 4.90E-03 1.00E+01 90th
: 4.90E-02 Haas et al. (1999); Soller et al. (2004)
Vibrio cholera 2.00E+00 1.54E-02 2.73E+00 95th
: 4.21E-02 Haas et al. (1999)
Yersinia spp. 1.00E+00 1.02E-03 1.00E+01 90th
: 1.02E-02 Lathem (2005)
Adenovirus 1.00E+00 4.17E-01 90th
: 1E+00 Haas et al. (1999)
Ascaris 1.00E+00 9.49E-02 1.00E+01 90th
: 1E+00 Mara and Sleigh (2010)
Coliphage
(Somatic) 0.00E+00 0.00E+00
111
Pathogen D-R Type Mean UF UB Reference
Enteroviruses 1.00E+00 2.00E-03
Regli et al. (1991) (for Echovirus12:
ingestion route); For inhalation: Tanner et
al. (2008) ; Brooks et al. (2005a,b) ; Haas
et al. (1999) (using Coxsackievirus B5
dose-response model)
Hepatitis A virus 1.00E+00 5.49E-01 90th
: 19E+00 Haas et al. (1999)
Hepatitis E 1.00E+00 1.30E-02 1.00E+01 90th
: 1.30E-01 Bouwknegt et al. (2009)
Astrovirus 1.00E+00 6.06E-07 1.00E+01 90th
: 6.06E-06 Commission on Life Sciences et al (2000)
Legionella 1.00E+00 6.00E-02 1.00E+01 90th
: 6.00E-01 Armstrong and Haas (2007)
Norovirus 1.00E+00 2.78E-04 Teunis et al. (2008)
Rotavirus 1.00E+00 6.19E-01 90th
: 1E+00 Regli et al. (1991); Haas et al (1999)
Toxoplasma 0.00E+00 0.00E+00
Fecal coliforms 0.00E+00 0.00E+00
E-coli 0.00E+00 0.00E+00
Enterococci 0.00E+00 0.00E+00
112
Table A-5 Inputs describing site characteristics and application events (Chapter 3)
Parameters Values Units Description
/Remarks
Rainfall intensity 0.25-0.33 cm/h Water application intensity and duration
were assumed to be the averaged values:
0.29 cm/h and 90 hours Rainfall duration 72-108 h
Soil texture class Sandy loam -
Area of application
site
20 Acre Assume a square plot
Pathogens and indicators in the biosolids (2008; L1-L3)
Potassium chloride 9×103 ppm 9 g/L (one mole in 4L water=(36 g)/(4L)),
with assumed 1g/mL density of biosolids.
P-22 1.5×108 PFU/g
biosolids
3×1011
PFU/100mL in biosolids with
assuming 1g/mL density and solid
percentage of 5%.
Somatic phage NA PFU/g
biosolids
Not detected in biosolids.
Adenovirus 2.1×105 PFU/g dry
biosolids
4.20×108 PFU/100mL in biosolids with
assumed 1g/mL density and solid
percentage of 5%.
Pathogens and indicators in the biosolids (2009; L4-L6)
Potassium chloride 9×103 ppm 9 g/L (one mole in 4L water=(36 g)/(4L)),
with assuming 1g/mL density of
biosolids.
P-22 7.5×106 PFU/g
biosolids
1.25×1010
PFU/100mL in biosolids with
assumed 1g/mL density and solid
percentage of 6%.
Somatic phage 48 PFU/g
biosolids
8×104 PFU/100mL in biosolids with
assumed 1g/mL density and solid
percentage of 6%.
Adenovirus 1.98×104 PFU/g dry
biosolids
3.30×107 PFU/100mL in biosolids with
assumed 1g/mL density and solid
percentage of 6%.
113
Table A-6a Model inputs with uncertainties: microbial parameters (Chapter 5)
Pathway Parameter Distrib
ution
type
Source Adenovir
uses
Cryptosp
oridium
Enterovir
uses
Giardia
lamblia
Salmonell
a spp.
Shigella
spp.
Surface
water,
groundwat
er, air, soil,
vegetables
Microbial
initial
concentration
in biosolids
(number/L)
log-
normal
Pepper et al.,
2010; Guzman,
2007
1.76×101(
3.70,
2.26×101)
2.80×101(
1.30×101,
6.40×101)
1.05×10-
1(2.50×10
-21,
8.00×10-
1)
(1.28×101
,
2.00×10-
19)
8.10×10-
1(2.50×10
-21, 3.35)
4.49(1.00
×10-20
,
9.20)
Surface
water,
groundwat
er, air, soil,
vegetables
Microbial
release factor
log-
normal
Chetochine et al.,
2006; Xagoraraki,
2010
3.40×10-
2(1.00×10
-2,
7.40×10-
2)
3.40×10-
2(1.00×10
-2,
7.40×10-
2)
3.40×10-
2(1.00×10
-2,
7.40×10-
2)
3.40×10-
2(1.00×10
-2,
7.40×10-
2)
3.40×10-
2(1.00×10
-2,
7.40×10-
2)
3.40×10-
2(1.00×10
-2,
7.40×10-
2)
Surface
water,
groundwat
er
Decay for
microbes in
water
(log/hour)
log-
normal
Enriquez, 1995;
Medema,
1998;Lyon and
Chattopadhyay,
2001; John et al.,
2005; Henis, 1987
1.75×10-
3(N/A,
1.75×10-
2)
1.58×10-
3(N/A,
1.58×10-
2)
5.00×10-
3(1.18×10
-3,
6.00×10-
2)
3.80×10-
3(N/A,
3.80×10-
2)
9.59×10-
3(2.88×10
-3,
5.75×10-
2)
4.40×10-
3(N/A,
4.40×10-
2)
Soil,
vegetables
Decay for
microbes in
soil (log/hour)
log-
normal
Enriquez, 1995;
Medema,
1998;Lyon and
Chattopadhyay,
2001; John et al.,
2005; Henis, 1987
1.75×10-
3(N/A,
1.75×10-
2)a
1.58×10-
3(N/A,
1.58×10-
2)a
5.00×10-
3(1.18×10
-3,
6.00×10-
2)a
3.80×10-
3(N/A,
3.80×10-
2)a
9.59×10-
3(2.88×10
-3,
5.75×10-
2)a
4.40×10-
3(N/A,
4.40×10-
2)a
Air Decay for
microbes in air
(log/day)
log-
normal
Harper, 1961;
Benbough, 1971;
Cox, 1968
1.00(9.80
×10-2
,
1.92)b
N/Ac 1.00(9.80
×10-2
,
1.92)
N/Ac 3.11
(9.70×10-
1, 5.24)
d
3.11
(9.70×10-
1, 5.24)
d
114
Pathway Parameter Distrib
ution
type
Source Adenovir
uses
Cryptosp
oridium
Enterovir
uses
Giardia
lamblia
Salmonell
a spp.
Shigella
spp.
Air Pathogen
inhalation-
related dose-
response
models
log-
normal
Haas et al., 1999;
Soller et al., 2004;
Regli et al., 1991;
Tanner et al.,
2008; Brooks et
al., 2005c;
4.17×10-
1(4.17×1
0-2
, 1.00)
N/Ac 2.53×10
-
2(2.53×10
-3, 1.00)
N/Ac 1.36×10
-
6(3.55×10
-7,
1.55×10-
4)
2.45×10-
3(2.45×10
-4,
2.45×10-
2)
Surface
water,
groundwat
er, soil,
vegetables
Pathogen
ingestion-
related dose-
response
models
log-
normal
Haas et al., 1999;
Soller et al., 2004;
Regli et al., 1991;
Tanner et al.,
2008; Brooks et
al., 2005c;
4.17×10-
1(4.17×1
0-2
, 1.00)
4.19×10-
3(2.15×10
-3,
7.57×10-
3)
2.00×10-
3(2.00×10
-4,
2.00×10-
2)
2.00×10-
2(4.40×10
-3,
5.66×10-
2)
2.71×10-
6(7.10×10
-7,
3.09×10-
4)
4.90×10-
3(4.90×10
-4,
4.90×10-
2)
Groundwat
er
Microbial
radius (cm)
uniform Dai, 2006;
Hayhow, 1993
4.00×10-
6(3.50×10
-6,
4.50×10-
6)
3.30×10-
4(2.75×10
-4,
3.85×10-
4)
1.18×10-6
5.90×10-
4(5.25×10
-4,
6.55×10-
4)
5.75×10-5
5.75×10-5
Note: Values displayed are averaged values with 5-95th percentiles given in parentheses. Italic values represent values by assuming 10 as
uncertainty factor aDecay rates in soil are assumed to be same as decay rates in water (recommended to be same by Lyon 2001, and
observed similar for MS2 and PRD1 by Anders,2009 and Torkzaban,2006) bDecay rates in air for adenoviruses are assumed to be same as enteroviruses
cParasites are not an issue in aerosols, so the decay rates and dose-response model are not available
dDecay rates in air for Salmonella spp. And Shigella spp. are assumed to be same as E.coli
115
Table A-6b Model inputs with uncertainties: soil parameters (for sandy-loam soil only)
(Chapter 5)
Pathway Parameter Value Distribution
type
Source
Groundwater Hydraulic
conductivity
(cm/h)
4.42(0,
1.01×101)
normal Carsel and
Parrish,
1988
Surface
water,
groundwater
Saturated
water content
4.10×10-
1(3.20×10
-1,
5.00×10-1
)
uniform Carsel and
Parrish,
1988
Surface
water,
groundwater
Residual
water content
6.50×10-
2(4.8×10
-2,
8.2×10-2
)
uniform Carsel and
Parrish,
1988
Groundwater Effective
dispersion
factor
1.00(1.00,
2.31)
uniform Teng et al.,
2012b
Groundwater Retardation 3.16(1.60×10-
1, 3.46)
uniform Teng et al.,
2012b
Note: Values displayed are averaged values with 5-95th percentiles given in
parentheses.
116
VITA
Jingjie Teng was born in Jiangsu, China on April 9, 1984. She received her
bachelor’s degree of Environmental Engineering from Beihang University, Beijing,
China in 2006. She became a Master of Science in Environmental Engineering at
Drexel University in 2006. She earned the Doctoral degree in Environmental
Engineering in 2012. She has been rewarded and published multiple works during her
Ph.D study on quantitative microbial risk assessment and subsurface fate and
transport modeling.
Awards
Student Award for presentation “Microbial risk assessment of exposure to biosolids-
associated pathogens” by Society for Risk Analysis, Baltimore, MD, December 2009.
Student Podium Award Winner for presentation "Extending the risk assessment
framework for pathogens in biosolids: Groundwater Pathway" by Pennsylvania Water
Environment Association, Lancaster, PA, June 2009.
Selected Pulications
Teng, J., A. Kumar, H. Zhang, M. S. Olson, and P. L. Gurian, (2012)
“Determination of Critical Rainfall Events for Quantitative Microbial Risk
Assessment of Land-Applied Soil Amendments”, Journal of Hydrologic Engineering,
17(3): 437-444.
Teng, J., A. Kumar, P. L. Gurian, M. S. Olson, and H. Zhang. (2010)
“Determination of Critical Rainfall Events for Quantitative Microbial Risk
Assessment of Biosolids-Associated Pathogens”, WEF Residuals and Biosolids
Conference, May 23-26, 2010, Savannah, GA.
Teng, J., P. L. Gurian, M. S. Olson, A. Kumar, H. Zhang, C. Harte, B. Olson, and K.
Downs. (2009) “Microbial risk assessment of exposure to biosolids-associated
pathogens”, Society for Risk Analysis Annual Meeting, December 7, 2009, Baltimore,
MD.
Teng, J., M. S. Olson, and P. L. Gurian. (2009) “Extending the risk assessment
framework for pathogens in biosolids: Groundwater pathway”, Pennsylvania Water
Environment Association Meeting, June 2009, Lancaster, PA.
