Associate;. Inc.
Human Health Risk
Assessment for the
Use and Disposal of
Sewage Sludge: :
Benefits of Regulation
Prepared for:
U.S. Environmental Protection Agency
Office of Water
Office of Science an
-------
Acknowledgments
* *u ™s ^r^P1^31^^^ Associates, toe., under Contact Number 68-CO-0093
Sf?e Vw ^STSy1 ^^^ A^W8 Health Md Ecological Criteria Division of the
Office of Water. The following Abt Associates staff were the major contributors to this analysis
and document: J
KirkmanONeal Principal Investigator
Vicfa Hutson Project M-magg,.
Michael Conti Advisor/Technical Reviewer
Dan McMartin Environmental Modeler
Elizabeth Fechner Levy Environmental Scientist
Wendy Hughes Environmental Modeler
Susan Keane Environmental Scientist
Kathy Cunningham Toxicologist
11311 Wa°g Research Assistant
Barry Lester Consultant
Abt Associates staff would like to thank Dr. Alan Rubin for his guidance and support as EPA
Project Manager. We would also like to thank Robert Southworth, Mark Morris, Norma
Whetzel, Barbara Corcoran, and Gene Grumpier of the Office of Water for their useful
comments and valuable insights on various aspects of this study.
-------
Table A
Estimated Baseline Cancer Risks from Sewage Sludge
Aggregate
Cancer Risk
(cases/yr)*
Risk to Highly
Exposed
Individual
(risk/lifetime)*
Incineration1"
Risk to Average
Exposed
Individual
(risk/lifetime)*
0.3-4
6X104 - 7xlO-3 2xlO-7 - 3x10-*
0.5
0.07
0:9^5
6xlO-»
IxlO-7
2x10^
Land Application
Surface Disposal
All Praedees-Combined'''c
' AH values independently rounded to one sigiuint figure. Values may not sum to
totals because of independent:, rounding. y
1JCT ^ l?Sed °n "**?' estimate" of emissions for organic pollutants; higher
value based on "worst case" estimate of emissions fonorganic pollutants
Reports total in aggregate risk column, maximum in column for risks to HEI and
combined average for average exposed individual, column (see text for further '
- 7x10*' 3xTO-7~-
-------
c
(0
o
Q.
GQ
^
<.<2
©
O) JL.
E o
O
c
(0
O
0)
£
"3
0)
CO
QQ
o
o
Q.
g
"c
O)
o
co
c
g
"w
c/
o
cd
CD
CD Q.
'o w
y CD
.£ co
M_ CO
o o
is
CO T3
CD CD
P td
c E
0 to
-O CD
CO CC
03 (^
CD £
-------
this analysis. ^Exposure to arsenlc^is responsible for less than one percent of total risks through
inhalation and dietary pathways, but dominates estimated cancer risks from the groundwater and
direct ingestion pathways, ultimately contributing more than 20 percent of total estimated cancer
risks from all pathways.{Arsenic was detected in 80 percent of sludge sampled for the NSSS.
The next largest contributors to estimated cancer risks are aldrin and dieldrin, which were
detected in 8 percent of samples' for the NSSS and contributed 17% of total cancer risk.
Exposure to aldrin/dieldrin is estimated to occur primarily1 through emissions from incinerators.
Neither of these compounds has been detected in emissions from incinerators; as with
hexachlorobenzene, our estimates are based on limits of detection. PCBs were detected in 19
percent of samples from the NSSS, and account for about 11 percent of estimated cancer risk.
More than 90 percent of the cancer risk estimated for PCBs occurs through the dietary pathway,
for which direct ingestion by grazing animals can lead to human exposure. Exposure to
cadmium (which was found in 69 percent of samples) is responsible for about 8 percent of
estimated cancer risk. All estimated cancer risk from cadmium occurs through the inhalation
pathway from incinerators. Most of the remainder of the estimated cancer risk is caused by
products of incomplete combustion (PICs) emitted by incinerators. This analysis includes the
modeling of 85 organic pollutants, most of them PICs, emitted by incinerators.
For non-cancer health effects, we compare estimated exposure to available risk reference
doses (RfD). In all cases, average lifetime exposure for each of the human populations modeled
falls beneath the risk reference dose for all contaminants considered. For two management
practices, however, hypothetical (reasonable worst case) exposure scenarios for the HEX suggest
that exposure to arsenic might exceed risk reference doses under baseline conditions. Numerical
criteria in the regulation, which iare based on estimated exposure for the HEI, are expected to
eliminate this condition. For lead and cadmium, we use additional calculations to determine
potential health effects. We find that fewer than one, individual per year is expected to exceed
a threshold concentration of cadmium in the kidney cortex through exposure to cadmium from
sludge. For lead, Table B summarizes our estimates of the number of persons crossing blood
lead thresholds (7 ng/dl for men and 10 /tg/d/ for women and children) because of exposure to
sludge. We estimate that about 2000 persons per year could cross these thresholds because
exposure to sludge; of these, about 70 percent are due to exposure to land application. Included
in this total are children who cross blood lead thresholds because they ingest small amounts of
garden soil contaminated by sludge used for home gardening. Table B also contains estimates
for the number of cases of lead-related disease caused by exposure to sludge. Health endpoints
evaluated include cases of IQ reduced below 70 for children; and hypertension, stroke,
cardiovascular disease and death in adult men. About 80 percent of the estimated cases are
caused by land application of sludge, mostly through residential uses. Because of the relative
immobility of lead in soil, estimated risks from exposure to lead from surface disposal are quite
low.
Under current, baseline conditions, more than half of the sludge represented by the NSSS
is applied to land, with about 30 percent incinerated and about 15 percent managed with surface
disposal. Because regulation could possibly lead to the shifting of sludge from one practice to
another it is useful to consider the .relative risk of alternative management practices for sludge.
Table C compares risks from ea,ch type of management practice per million tons of sludge
managed. Based on the "best" estimate of emissions from incinerators, cancer risks per million
-------
Table B
* •
Estimated *Risks from Exposure to Lead from Sewage Sludge
Expected Cases of
Persons Crossing Blood ' Health Effects*-6
_ _ Lead Thresholds*'1* (cases/yr)
Incineration 700,
Land Application 1000 500
Surface Disposal . . <1 : .
-------
Table C
«
I
Comparison of Baseline Risks per Million Metric Tons of Sewage Sludge
For Alternative Management Practices
Persons Cases of
Crossing Lead Disease
Canceir*;b Threshold1'6 from LeadM
(cases/mint) (cases/mint) (cases/mint)
Incineration
0.4-5
800
100
Land Application
Agriculture: Dietary Pathways
Groundwater/Surface Water/Air
Residential Uses
Total
0.4
0.07
0.08
0.3
300
10,000
1000
20
<1
4000
400
Surface Disposal
Monofills 0.1
Surface Impoundments and Other 0.2
Total 0.2
All Practices Combined
0.3-2
800
200
* All values reflect expected health effects per million metric tons (mint) of sludge
managed, and are independently rounded to one significant figure. Values may not
sum to totals because of independent rounding.
b Incremental cases of cancer per year from exposure to pollutants from sludge.
Where a range is provided, the lower value is based on "best estimates" of emissions
for organic pollutants from incineration. The higher value is based on "worst case"
estimates of emissions.
0 Number of persons crossing thresholds for concentration of lead in blood: 7 /tg/dl
for men, and 10 /*g/dl for women and children.
d Includes hypertension, stroke, cardiovascular disease, and death for men; cases of
IQ reduced below 70 for children.
Vll
-------
tons are of comparable magnitude for all three groups of practices considered. Within these
groups, residential uses for land-applied sludge possibly offer the lowest cancer risk per ton,
followed by the disposal of, sludge in monofills. I
Estimated risks from lead follow a different pattern. As shown by the second and third
columns of Table C, risks per ton are highest for the residential uses of sludge, where potential
direct ingestion of contaminated soil by children presents significant estimated risks. Estimated
risks from lead are lowest for surface disposal.
Health Benefits from Regulation
In response to the regulation, treatment works operating incinerators are expected to
retrofit their facilities with additional pollution control devices, or to operate them at higher
efficiency. These additional contools are expected to achieve significant reductions in emissions
and resulting cancer risks, as shown in Table D. The regulation's impact on risks from land
application and surface disposal could not be determined with available data and the methods
used for this analysis; however, potential, reductions in (risk should not exceed the baseline
estimates listed in the table. ,
Similarly, Table E shows how the regulation is expected to reduce health risks from
exposure to lead. Installation of electrostatic precipitators at some incinerators is expected to
eliminate about 80 percent of the cases of lead-related disease estimated for baseline conditions,
and about 90 percent of the estimated number of persons crossing thresholds in blood lead.
Depending on its effects on land application and surface disposal, the regulation is expected to
result in the avoidance of 600-2000 persons crossing blood lead thresholds, and 90-600 cases of
lead-related disease per year.
Conclusions
Results from this study should be viewed in light iof the limitations to be discussed in
Chapter 1. These results should serve as a useful input to the regulatory decision-making
process, but further research is needed to focus on a number of areas omitted from this study.
One such area is ecological effects, for which the Agency has not yet developed an accepted
methodology for assessing risks. Another is risks from pathogenic organisms. In practice, these
are a major determinant in choices of methods for sludge treatment or disposal. However,
methods are not currently available for providing useful quantitative estimates of risks from
pathogenic organisms. Additional pathways of human exposure to chemical pollutants should
also be considered. In particular, future research should assess risks from indirect pathways of
exposure to pollutants from incineration of sludge. Such research would assess risks from
deposition of pollutants onto surface water, soil, or crops. It would also consider potential
health risks from the disposal of incinerator ash. Additional technical research is needed to
quantify the uncertainty implicit in the estimates reported here, and would ideally provide
confidence intervals for all risk estimates.
viu
-------
Table
Expected Reduction in Cancer Risks
From Regulating Sewage Sludge*
Baseline Risk* |
Risk After
Regulation*
Reduction in
Risk*
INONERATION0
Aggregate Risk
Risk to AEI
Risk to HEI
0.3 -4
2xlO7 - 3x10*
6x10^ - 7xlO-3
0.2-4
IxlO-7 - 3x10*
4x10^ - 7xlO-3
0.09
2-28%
3-39%
LAND APPLICATION
Aggregate Risk
Risk to AEI
Risk to HEI
0.5
IxlO-7
6X104
0-0.5
0 - IxlO-7
o -
0-0.5
0-100%
0- 100%
SURFACE DISPOSAL
Aggregate Risk
Risk to AEI
Risk to HEI
0.07
2x10-*
6x10*
0 - 0.07
0 - 2x10*
0 - 6X1O4
0 - 0.07
0- 100%
0- 100%
ALL PRACTICES COMBINED'1
Aggregate Risk
Risk to AEI
Risk to HEI
0.9-5
3xlO"7 -
6x10-* - 7xlO3
0.2-4
6xlO* - 1x10*
4x10-* - 7xlO3
0.09 - 0.7
2-75%
3-39%
1 For aggregate risk, reports toital incremental cases of cancer expected per year. For risk
to average exposed individual (AEI) and highly exposed individual (HEI), reports
incremental individual risk of cancer from lifetime of exposure to pollutants from sludge.
All values are independently rounded to one significant figure. Values may not sum to
totals because of independent rounding.
b For aggregate risk, reports estimated number of canceri avoided per year. For risk to
AEI and HEI, reports percent reduction in risk, calculated as: '
(Baseline - Control)/Baseline.
c Lower value based on "best estimate" of emissions for organic pollutants; higher value
based on "worst case" estimate of emissions for organic pollutants.
d Reports total for aggregate risks, average for risk to AEI, and maximum for risks to
HEI.
IX
-------
Table E
Expected Reduction in Lead-Related Health Risks
From Regulating Sewage Sludge*
Risk After
Baseline Risk1 Regulation
INCINERATION1'
Persons Crossing Lead Threshold
Cases of Disease from Lead
LAND APPLICATION
Persons Crossing Lead Threshold
Cases of Disease from Lead
700
100
1000
500
: 90
30
0 - 1000
0-500
Reduction in
Risk
600
90
0- 1000
0-500
SURFACE DISPOSAL
Persons Crossing Lead Threshold
Cases of Disease from Lead
ALL PRACTICES COMBINED
Persons Crossing Lead Thnsshold 2000 20 - 2000 600 - 2000
Cases of Disease from Lead 700 " 30 - 600 90 - 600
* All values are independently rounded to one significant figure. Values may not sum
to totals because of independent rounding.
b Lower value based on "besit estimate" of emissions for organic pollutants; higher
value based on "worst case" estimate of emissions for organic pollutants.
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY
TABLE OF CONTENTS ....
LIST OF TABLES .......
LIST OF FIGURES
LIST OF MATHEMATICAL SYMBOLS
1. INTRODUCTION ................. ....... ...... ............ 1-1
1.1 GENERAL FRAMEWORK FOR RISK ASSESSMENT AND RISK-
BASED BENEFITS ANALYSIS^ ...... ............. ....... 1_1
1 .2 APPLICATION OF METHODOLOGY TO THE USE AND DISPOSAL
OF SEWAGE SLUDGE .......... _
1.2.1 Sludge Pollutants ........... \ ............... .... \.\2
1.2.2 Disposal Options .......... , ; ....... . ....... ... 1-14
1.2.3 Fate and Transport Modeling . . . . ................... 1-16
1.2.4 Exposure Pathways .......... ............ . . ..... i_ig
1.2.5 Populations Exposed ......... i ........ .... ....... 1-18
1.2.6 Health Effects ............. ..... .............. 1-18
1.2.7 Risk Characterization ........ ................... 1-22
1.2.8 Benefits Analysis and Compliance Strategies ............ 1-23
1.2.9 Limitations/Uncertainty ....... ................... 1-25
2. LEAD AND CADMIUM ................. | ........... .... ..... 2-1
2.1 ESTIMATING HEALTH EFFECTS FROM LEAD EXPOSURE . . ... 2-1
2.1.1 Background Exposure ........ : .................... 2-1
2.1.2 Absorption and Uptake of Lead ..'... .................. 2-3
2.1.3 Dietary or Drinking Water Pathways ................... 2-7
2.1.4 Shifting the Blood Lead Distributions ................... 2-7
2.1.5 Health Effects ............. .................... 2-9
2.2 ESTIMATING HEALTH EFFECTS FROM CADMIUM .... ...... 2-18
xi
-------
3. INCINERATION ...... ................
3.0 INTRODUCTION ... ' , ,
* ........ • ......... ............ 3-1
3.1 METHODOLOGY
3.1.1 Estimating; Emissions of Pollutants ............ 3_2
3.1.2 Modeling the. Dispersion of Pollutants in Air ........ .... 3-3
3.1.3 Mapping Dispersion and Pollutant Concentrations Onto a Unified
Grid ......... .............................. 34
3.1.4 Estimating Human Exposure and Risk ............. 3_g
3.1.5 Estimating Benefits from Regulation ..'..'.'.'.'.'.'.','.'.'.'.'.'.'. 3-H
3.2 DATA SOURCES! AND MODEL INPUTS ....... 3_!2
3.2.1 National Sewage Sludge Survey . . ............. 3_j2
3.2.2 Other Sources of Facility-Specific Data ........ . . ... 3-12
3.2.3 Furnace Types .......... . ......... .... ........ 3.13
3.2.4 Air Pollution Control Devices . . ., .......... ..... 3-13
3.2.5 Current Inventory of Furnaces and Pollution Controls ...... . . 3-16
3.2.6 Expected Response to Regulatory Controls ........ ...... 3-16
3.2.7 Data for Estimating Emissions of Inorganic Pollutants .... . . . 3-18
3.2.8 Emissions of Organic Pollutants . ; ...... ...... 3.31
3.2.9 Population Data ...... ...... ; ...... .......... 3-31
3.2.10 Meteorological Data ......... . ........ 3 3g
3.2.11 Health Effects Data .......... ;. . . ] .' .........I''' 3.35
3.3 RESULTS AND DISCUSSION ....... ......... .......... 3.36
3.3.1 Baseline Risks ............. ........... 3_3g
3.3.2 Benefits from Regulatory ^Controls . ............... 3.45
4. LAND APPLICATION: DIETARY PATHWAYS ... .................... 4_!
4.0 INTRODUCTION ............... ! 4_!
4. 1 METHODOLOGY .............. ......... 4_j
4.1.1 Overview . ..... ..... ..... ; ............ 4-1
4.1.2 National Versus Local Aggregation . .............. .4-2
4.1.3 Description of Calculations ..... ; ............ ...... .4-3
4.2
DATA SOURCES AND MODEL INPUTS . •.' 4.7
4.2.1 Application Rates r 4_jj
4.2.2 Concentration of Contaminants in Sludge 4-11
xu
-------
4.3
4.2.3 Uptake Rates
4.2.4 Animal Feed Mixes ...............
4.2.5 Dietary Consumption ........ . . '. ............... 4^7
4.2.6 Fraction of Crops Grown in Sludge-Amended Soil . . . . . . . . . . 4.17
RESULTS AND DISCUSSION . .....; ............. . 4_21
4.3.1 Baseline Risks: Dietary Pathway , .............. 4_2j
4.3.2 Benefits from Regulatory Controls : .......... ......... 4-26
5. LAND APPLICATION: GROUNDWATER, SURFACE WATER AND AIR
Jr A JL jj. iVA x S ...... , « ..... #> ci
5.0 INTRODUCTION .............. ....... 5 j
5.1 METHODOLOGY . . . .......... 5 j
5.1.1 Mass Balance .......... . . j ..... 5_,
5.1.2 Estimating the Concentration of Contaminants in Groundwater . 5-5
5. 1 .3 Estimating the Concentration of Contaminants in Ambient Air 5-8
5.1.4 Estimating the Concentration of Contaminants in Surface Water ." 5-10
5.1.5 Estimating Human Exposure and Risks ............... S_15
5.2 DATA SOURCES AND MODEL INPUTS ...... 5_17
5.2.1 Sludge Parameters and Site Parameters ........... 5_17
5.2.2 Soil Parameters . ...- ......... .......... ......... 5.21
5.2.3 Hydrologic Parameters . . ...... ..... ...... ........ 5.25
5.2.4 Chemical-Specific Parameters * ' ' ' ' '
... . . . . . . . . . .
5.2.5 Size of Exposed Population ..'...-...'....'.','.'.', ....... 5.
33
5.3 RESULTS AND DISCUSSION ....... ............. 5.42
5.3. 1 Baseline Risks: Groundwater, Surface Water, and Air Pathways . 5-42
5.3.2 Benefits from Regulatory Controls . ............. 5_42
6. LAND APPLICATION: RESIDENTIAL USES
6.0 INTRODUCTION .....
.6-1
6.1 METHODOLOGY . ,- ,
• • 6-1
6.1.1 Overview , ; ,-_,
6.1.2 Description of Calculations 6-2
xm
-------
6.2 DATA SOURCES, AND MODEL INPUTS 6-7
6.2.1 Volume of Sludge to Home Gardens ' 6-7
6.2.2 Application Rates , 6-7
6.2.3 Concentrations of Contaminanits in Sludge . . 6-7
6.2.4 Uptake Rates for Crops 1 6-10
6.2.5 Dietary Assumptions ' 6-10
6.2.6 Exposed Population '. . . . . . ' 6-10
6.3 RESULTS AND DISCUSSION 6-15
6.3.1 Baseline Risks { 6-15
6.3.2 Benefits from Regulatory Controls • 6-20
7. SURFACE DISPOSAL 7_j
7.0 INTRODUCTION 7_j
7.1 METHODOLOGY 7_2
7.1.1 Algorithms for the Monofill Prototype 7.3
7.1.2 Methodology for Surface Impoundment Prototype 7-14
7.1.3 Estimating Human Exposure and Rislcs 7.25
7.2 DATA SOURCES AND MODEL INPUTS: .......... 7.26
7.2.1 Site and Sludge Parameters 7-27
7.2.2 Soil and Hydrologic Parameters 7.32
7.2.3 Chemical-S]pecific Parameters 7-40
7.2.4 Size of Exposed Population ', .....!] 7-51
7.3 BASELINE RISKS : •". 7.51
7.3.1 Benefits from Regulatory Controls .... 7-61
7.3.2 Uncertainties and Limitations .... 1 7.67
8. REFERENCES
APPENDIX A: Partitioning of Contaminant Among Air, Water, and Solids in Soil . . A-l
APPENDIX B: Derivation of First-Order Coefficient for Losses to Leaching ...... B-l
xiv
-------
APPENDIX C: Calculation of a TSquare Wave" for the Groundwater Pathway C-l
j.
C.I Land Application CM
C.2 Surface Disposal: Monofill Prototype C-2
C.3 Surface Disposal: Surface Impoundment Prototype C-3
xv
-------
LIST OF TABLES
Table A
Table B
Table C
Table D
Table E
Table 1-1
Table 1-2
Table 1-3
Table 1-4
Table 1-5
Table 1-6
Table 1-7
Table 2-1
Table 2-2
Table 2-3
Table 2-4
Table 2-5
Table 3-1
Table 3-2
Estimated Baseline Cancer Risks from Sewage Sludge
Estimated Risks from Exposure to Laid from Sewage Sludge
Comparison of Baseline Risks per Million Metric Tons of Sewage Sludge
for Alternative Management Practices . . .'. .". . .
x ! . • •
Expected Reduction in Cancer Risks from Regulating Sewage Sludge
Expected Reduction in Lead-Related health Risks from Regulating Sewage
Sludge 5
Classification Scheme for Weight of Evidence for Carcinogenicity
Sludge Pollutants by Management Practice ....
Quantity of Sludges by Management Practice
Models Used to Simulate Transport of Pollutants
Exposed Populations by Management Practices
Health Effects Data for Organic Contaminants in Sewage Sludge ....
Health Effects Data for Metals in Sewage Sludge
*. i
Estimated Intake Slopes: Increment in Blood Lead Concentration per Unit
of Exposure
Sample Calculation: Lead in Drinking Water . .
Potential Health Benefits from Reducing Exposure to Lead
Logistic Regression Relating Blood Pressure to the Probability of Initial
Cerebrovascular Accident in White Men Aged 45-74
Logistic Regression Relating Blood Pressure to the Probability of Initial
Atherothrombotic Brain Infarction in White Men Aged 45-74
Mass of Sludge Incinerated and Number of Incinerators by State ....
Furnaces and Pollution Control Devices for Incinerators in Analytic
Survey Before and After Regulatory Controls
m".«
vi
vii
ix
1-4
1-13
1-15
1-17
1-19
1-20
1-21
2-4
2-8
2-10
2-16
2-16
3-14
3-17
XVI
-------
Table 3-3 Metal Concentrations in Sludge Incinerated at Plants in Analytic Survey 3-19
Table 3-4 Metal Concentrations by Flpw Group for Sludge Incinerated at Plants in
the Analytic* Survey 3_2Q
Table 3-5 Combined Removal Efficiencies for Multiple Hearth Incinerators with Wet
Scrubbers ,' 3_23
Table 3-6 Combined Removal Efficiencies for Fluidized Bed Incinerators with Wet
Scrubbers . . . 3_24
Table 3-7 Summary of Removal Efficiencies for Metals . 3-25
Table 3-8 Removal Efficiencies of Pollution Control Devices: Summary of
Studies ; 3_2g
Table 3-9 Unit Emissions of Organic Pollutants 3-32
Table 3-10 Health Effects Data for Pollutants from Incineration of Sewage Sludge . 3-37
Table 3-11 Baseline Cancer Risk by Pollutant for Incineration of Sewage Sludge . . 3-42
Table 3-12 Contribution of Individual Pollutants to Total Estimated Cancer Risk from
Incineration . . 3.44
Table 3-13 Comparison of Exposure to Risk Reference Doses for Organic Pollutants
from Incineration of Sewage Sludge . . . 3-47
Table 3-14 Comparison of Exposure to Risk Reference Doses for Metals from
Incineration of Sewage Sludge .......: 3-48
Table 3-15 Cancer Risks from Incineration Before and After Regulatory Controls for
Incineration of Sewage Sludge ; . . . 3-49
Table 3-16 Risks from Lead jind Cadmium for Incineration of Sewage Sludge . . . 3-50
Table 4-1 Major Assumptions for Land Application: Dietary Pathway 4-8
>.
Table 4-2 Pollutant Concentrations in Sludge and Soil! 4-12
Table 4-3 Uptake Rates into Plant Tissue from land Application of Sludge .... 4-14
Table 4-4 Uptake Rates into Meat and Dairy Products from Animal Feed ..... 4-15
Table 4-5 Animal Feed Mixes 4-16
xvii
-------
Table 4-6
Table 4-7
Table 4-8
Table 4-9
Table 4-10
Table 4-11
Table 5-1
Table 5-2
Table 5-3
Table 5-4
Table 5-5
Table 5-6
Table 5-7
Table 5-8
Table 5-9
Table 5-10
Table 5-11
Table 5-12
Table 5-13
Table 5-14
Estimated Feed Mjx for Each Animal Product
Dietary Assumptions
Baseline Cancer Risks for Land Application: Dietary Pathway
Comparison of Baseline Exposure to Risk Reference Doses for Land
Application: Dietary Pathways
Comparison of Baseline Exposure of the! Highly Exposed Individual to
Risk Reference IDoses for Land Application: Dietary Pathways
Baseline Risks from Lead and Cadmium ;for Land Application: Dietary
Pathways ....;,
Parameters Used to Calculate az Under Stable Conditions
Pollutant Concentrations in Land-Applied Sludge
Site and Sludge IParameters for Land Application
Soil and Hydrologic Parameters for Land Application . .
Octanol-Water and Cyanic Carbon Partition Coefficients for Organic
Contaminants !
Octanol-Water Piittition Coefficients for PCBs
Distribution Coefficients for Organic and Inorganic Contaminants ....
x
Statistical Parameters for Predicting the Equilibrium Partitioning of Metals
in Surface Water ' . . .
Degradation Rates ; ;
Henry's Law Constants ;
Diffusion Coefficients for Contaminants in Air ...
Bioconcentration Factors and Food Chain Multipliers
State Population Densities
Baseline Cancer Bisks for Land Application: Groundwater, Surface Water,
and Air Pathways
4-18
4-19
4-22
4-23
4-24
4-25
5-11
5-18
5-19
5-22
5-27
5-28
5-30
5-31
5-33
5-34
5-36
5-37
5-41
5-43
XVIll
-------
Table 5-15
Table 5-16
Table 6-1
Table 6-2
Table 6-3
Table 6-4
Table 6-5
Table 6-6
Table 6-7
Table 6-8
Table 6-9
Table 7-1
Table 7-2
Table 7-3
Table 7-4
Table 7-5
Table 7-6
Table 7-7
Table 7-8
*.
Comparison of Baseline Exposure to Risk Reference Doses for Land
Application: Groundwater, Surface Water, and Air Pathways
«
Comparison' of Baseline Exposure for the HEI to Risk Reference Doses
for Land Application: Groundwater and Air Pathways ...
Major Assumptions and Their Ramifications for Land Application:
Residential Uses
Pollutant Concentrations in Sludge and Soil
Uptake Rates into Plant Tissue from Sludge Applied to Home Gardens
Dietary Assumptions for Land Application:' Residential Uses
Estimated Population Affected by Residential Uses of Sludge
Baseline Cancer Itisks for Land Application: Residential Uses
j. . ,
Comparison of Baseline Exposure to Risk Reference Doses for Land
Application: Residential Uses . .... . . '•'
Comparison of Baseline Exposure of the Highly Exposed Individual to
Risk Reference Doses for Land Application: Residential Uses
Baseline Noncancer Health Risks for Land Application: Residential
Uses
Parameters used to Calculate
-------
Table 7-9 Octanol-Water Partition Coefficients for PCBs ................ 7-44
Table 7-10 Degradation. Rates .............. ................... 7.45
t ;
Table 7-11 Henry's Law Constants ........... , ...... ............ 7.47
Table 7-12 Diffusion Coefficients for Organic Contaminants .............. 7-49
Table 7-13 Molecular Weights for Organic Contaminants ................ 7-50
Table 7-14 Data for Calculating Sizes of Exposed Populations for Monofills .... 7-53
Table 7-15 Estimated Sizes of Exposed Populations fojr Monofills ........... 7-54
Table 7-16 Data for Calculating Sizes of Exposed Populations for Surface
Impoundments .......... . ..... j. .................. 7^55
Table 7-17 Estimated Sizes of Exposed Populations fojr Surface Impoundments . . . 7-56
Table 7-18^ 'Baselne Cancer IU^^^ * . . . . . . 7-57
Table 7-19 Comparison of Baseline Exposure to Risk Reference Doses for Monofills:
Groundwater and Air Pathways ...... . .................. 7-58
I „ ...-."
Table 7-20 Comparison of Baseline Exposure for HEI to Risk Reference Doses for
Monofills: Groundwater and Air Pathways! ..... ............. 7-59
Table 7-21 Baseline Non-Caracer Health Risks: Monofilling ............... 7-60
Table 7-22 Baseline Cancer Cases: Surface Impoundments ... ............ 7-62
Table 7-23 Comparison of Baseline Exposure to Risk; Reference Doses for Surface
Impoundments: Groundwater and Air Pathways ............... 7-63
Table 7-24 Comparison of Baseline HEI Exposure to Risk Reference Doses for
Surface Impoundments: Groundwater and Air Pathways .......... 7-64
Table 7-25 Baseline Non-Caracer Health Risks: Surface impoundment ........ 7-65
Table 7-26 Baseline Cancer Cases: Total Surface Disposal ............... 7^66
xx
-------
LIST OF FIGURES
Figure A Baseline Cartcer ILisk by Pollutant iv
Figure 2-1 Probability of Hypertension vs. Blood Lead 2-12
Figure 3-1 Large Grid for Modeling Incinerators 3-6
Figure 3-2 Facility Location on Small Grid ......; 3-7
Figure 3-3 Population by Number of Overlapped Plumes, Incineration of Sewage
Sludge '. 3-41
Figure 3-4 Baseline Individual Cancer Risk by Population Size for Incineration . . 3-46
Figure 3-5 Increment to Blood Lead by Population Size for Incineration 3-52
Figure 5-1 Zones of Human Exposure for Groundwater, Surface Water, and Air
-- Pathways .-.,..-.•-,••...••=••-,;~. . . -..^v.—--.--iv-~-...~..v _v.v,v.>.. J; v.v.... 5.39-
Figure 7-1 Zones of Human Exposure for Groundwater and Air Pathways 7-52
xxi
-------
LIST OF MATHEMATICAL SYMBOLS
a
ABP
APr(CHD)
APr(HYP)
APr(MRT)
APr(STR)
P,.
Pw
ffz
*
V
T
a
A
ABLE
AF
empirical constant :
intermediate variable used to calculate emissions from contaminated soil
(m2/sec) .
longitudinal dispersivity (m) ;
empirical constant \
change in diastolic blood pressure (mm Hg)
"change in probability of CHD event (dimensionless)
change in probability of hypertension (dimensionless)
change in probability of death froiri all causes (dimensionless)
change in probability of stroke (dimensionless)
intermediate variable used to calculate volatile emissions from treated
soil
angle subtended by width of sludge management area or disposal site at
distance equal to estimated virtual distance from site (degrees)
air-filled porosity (dimensionless)
air-filled porosity in cover layer of soil (dimensionless)
total porosity in cover layer of soil (dimensionless)
effective porosity of soil (dimensionless)
total porosity (dimensionless)
water-filled porosity (dimensionless)
viscosity of air (g/cm-sec) ;
viscosity of water (g/cm-sec)
density of air (g/cm3)
particle density of sludge (kg/m3) '
particle density of soil (kg/m3) i
^particle density of sludge-soil mixture (kg/m3)
density of water (g/cm3 or kg//) :
standard deviation of the vertical distribution of contaminant in ambient
air (m) :
friction velocity of wind (m/sec)
empirically derived exponent (dimensionless)
effective porosity (dimensionless) :
pressure head (m)
air pressure head (m) : ••
first coefficient for calculating
-------
List of Mathematical Symbols (continued)
= applipation rate to SOT! for crop i of pollutant j (kg/ha, adjusted for
background soil concentration and for additional soil mass from added
sludge)
AR = application rate for sludge (Mg/ha-yr)
AR, = sludge application rate to crop / (Mg/ha-yr)
b = second coefficient for calculating az \
BAF = bioaccumulation factor (//kg)
BCF = contaminant-specific bioconcentration factor (//kg)
BD = bulk density of sludge/soil mix (kg/m3)
BIj = background intake of pollutant / (mg/kg-day)
BPi = diastolic blood pressure before change in exposure to lead (mm Hg)
BP2 = diastolic blood pressure after change in exposure to lead (mm Hg)
BSj - background soil concentration (dry wt) of pollutant j (mg/kg)
BW = average body weight (kg)
C = concentration of contaminant in sludge (mg/kg)
Q = concentration of contaminant in liquid layer of surface impoundment
(kg/m3) .
Qi = concentration of contaminant in sediment layer of surface impoundment
(kg/m3)
Q = concentration of contaminant in air-filled pore space of sludge/soil
mixture (kg/m3)
C.ir — average concentration of contaminant in ambient air at the receptor
location Otg/m3)
Cb = background concentration of contaminant in groundwater (mg//)
Cdw = dry weight concentration of contaminant in soil (mg/kg)
Q = concentration of contaminant in inflow to surface impoundment (kg/m3)
Cj = concentration of contaminant j in sludge (g/DMT)
Ci« = concentration of contaminant in water leaching from site (kg/m3 or
mg//) ;
C0 = concentration of contaminant in outflow from surface impoundment
(kg/m3)
C, = concentration of contaminant adsorbed to solids (kg/kg)
C«a = dry weight concentration of contaminant in eroded soil (mg/kg)
CM? = concentration of contaminant in seepage from surface impoundment
(mg//) *
Ctma = lthe concentration of contaminant in s0il eroding from the sludge
management area (mg/kg)
Csw = concentration of contaminant in surface water (mg//)
Q = total concentration of contaminant in sludge/soil mixture (kg/m3)
Cu = concentration of contaminant in unsaturated zone (g/m3)
Cw = concentration of contaminant in water-filled pore space of sludge/soil
mixture (kg/ra3)
Cwc( = concentration of contaminant in well-water (mg//)
CB = background concentration of contaminant in soil (mg/kg)
xxiii
-------
List of Mathematical Symbols (continued)
CDu = tissue concentration (dry wt) of pollutant y in crop i (mg/kg)
CF* ~ weighted average concentration of pollutant y across all food sources for
animal producing meat or dairy product k (jig/g)
CI = incremental cancer risk for exposed individual (incremental risk of
developing cancer per lifetime of exposure)
CIj •= incremental cancer risk from contaminant./ for exposed individual
(increment^ risk of developing cancer per lifetime of exposure)
C*m« = maximum detected total incremental cancer risk among all modeled
individuals (incremental risk of developing cancer per lifetime of
exposure) , •
cp = total aggregate cancer risk for exposed population (expected
incremental cases of cancer per year)
cpj = aggregate cancer risk from pollutant;/ (expected cases per year)
CR = contact rate; (mass per time or volume per time)
CTii = concentration of contaminant j in sludge-amended soil used to grow
crop /, adjusted for background soil concentration and for additional
soil mass for added sludge (mg/kg)
di = depth of liquid layer in surface impoundment (m)
di = depth of se = intermediate variable used to calculate emissions from contaminated soil
(nWsec)
Dr = dilution factor (dimensionless)
^V = rate of change of volume, positive for sediment layer in surface
impoundment, negative fqr liquid layer (mVsec)
EJP ~ emission rate for contaminant y at incineration facility p (grams/sec)
xxiv
-------
List of Mathematical Symbols; (continued)
ED = exposure duration (yr)
EXPjj •= exposure to contaminant j for persons in living in cell / (mg/kg-day)
EXPJ ~ incremental exposure to pollutant/ 'for a particular population
(mg/kg-day) *-*-
f« = fraction of total contaminant loading lost during active operation of
monofill facility (dimensionless) ;
f«« = fraction of annual loading of contaminant lost from surface
impoundment each year to all processes combined (dimensionless)
fco - fraction of facility's active lifetime that typical cell contains sludge and
temporary .soil cover (dimensionless)
fdi = fraction of contaminant in the liquid layer of surface impoundment that
is dissolved (dimensionless) ;
fd2 = fraction of contaminant in the sediment layer of surface impoundment
that is dissolved (dimensionless)
f
-------
List of Mathematical Symbols (continued)
f"p» = fraction of mass entering the sediment layer of surface impoundment
that fe lost to seepage (dimensionless)
f>1 = fraction of monofill's total volume containing pure sludge (m3/m3 or
dimensionless)
f«i = fraction of solids in sludge (kg/kg or dimensionless)
fua = fraction of monofill's active lifetime that typical cell contains uncovered
sludge (dimensionless)
f« = fraction of contaminant loss during, monofill's active lifetime that is lost
to volatilization (dimensionless) ;
fvi = fraction of contaminant loss from inactive monofill that is lost to
volatilization (dimensionless)
fv" = fraction of total contaminant loading to monofill that is lost to
volatilization during a time interval: equivalent to a human lifetime
(dimensionless)
fv°< = fraction of total contaminant loss caused by volatilization
(dimensionless)
fv»ii = fraction of contaminant leaving the liquid layer that leaves through
volatilization (dimensionless) •
fwd = ratio of contaminant concentration in well-water to concentration in
seepage beneath the surface disposal facility (dimensionless)
'LS = fraction of total cumulative loading lost in human lifetime
(dimensiomtless) . | • -
F> = volume of fluid passing through a vertical cross section of the aquifer
oriented perpendicular toxthe direction of flow, and having a width
equal to the source width and a depth equal to the saturated thickness of
the aquifer (m3/sec) j
F* = fraction of animal's food from crop./ for animal product k
(dimensionless) j
FA' ~ annual flux: of contaminant leaching Ifrom treated land (kg/ha-yr)
FAv = annual average flux of contaminant volatilizing from treated land
(kg/ha-yr) '
FS; = flux of leached contaminant from mixture of sludge and soil (g/m2-yr)
FCi = fraction of dietary consumption of crop i grown in sludge-amended soil
(dimensionless)
FCk = fraction of dietary consumption of animal product k produced with
sludge (dimensionless) i
FD ~ ratio of effective diameter to depth of surface unpoundment
(dimensionlless)
FM = food chain multiplier (dimensionless)
FS|= = direct ingestion of sludge (adherence pathway) as fraction of animal's
diet for animal producing food product k (dimensionless)
H = Henry's Law constant (m3-atm/mol)
H = Henry's Law constant (dimensionless at specified temperature)
> = index for grid cells * '.
xxvi
-------
List of Mathematical Symbols (continued)
i = index for crops
J. = inhalation volume (mVday)
*f - daily consumption of fish (kg/day)
J« = individual intake for persons living in cell / of combined emitted
contaminant j from all incinerators (mg/kg-day)
*«« = average rate of soil ingestion for adults (g/kg-day)
IK = average rate of soil ingestion for children (g/kg-day)
*t = daily dose of contaminant from fish! (mg/day)
Iw = rate of water ingestion (//day)
Ici = estimate of maximum individual cancer risk for contaminant j
(probability of developing cancer from lifetime dose)
J = index for contaminants
krw = effective permeability (dimensionless)
K = saturated hydraulic conductivity (m/sec)
= degradation rate coefficient for monofill or sludge management site
1
K
-------
Kvl
Kvi
KD
KDSW
KOC
KOW
-LA:—
List of Mathematical Symbols (continued)
rate coefficient for loss of contaminant to volatilization from active
monofill (yr1)
rate coefficient for loss of contaminant to volatilization from inactive
monofill (yr1)
rate coefficient for loss of contaminant to volatilization from treated
land (yr1) :
rate coefficient for loss of contaminant to volatilization from the liquid
layer of surface impoundment (m/sec)
equilibrium partition coefficient for contaminant (m3/kg)
partition coefficient for contaminant in the stream (//kg)
organic caibon partition coefficient (mVkg)
octanol-waiter partition coefficient fo;r contaminant
distance between the SMA and the receiving water body (m)
mass of coiiSminant added to sou treated with sludge (kg/ha)
mass of contaminant from non-sludge sources in treated soil (kg/ha)
active lifetime of monofill facility: the period in which the facility
accepts sludge (yr)
lifespan of average individual (yr)
mass of g&ieous contaminant (kg)
mass of adsorbed contaminant (Teg)
total mass of contaminant in soil (kg)
mass of dissolved contaminant (kg)
mass of contaminant in monofill after end of year LF (kg/ha)
mass of contaminant in treated soil at end of period equal to average
human lifespan (kg/ha)
mass of contaminant in soil after N applications (kg/ha)
mass of soil (kg)
mass of sludge incinerated each year at incinerator/? (DMT/yr)
mass of contaminant in treated soil at time / (kg/ha)
Maximum Contaminant Level for drinking water (mg/0
mass of soil eroding annually (i.e., rate of soil loss) from one hectare
of sludge management area (kg/ha-yr)
estimated rate of soil loss for the watershed (kg/ha-yr)
mass of solids in one m3 of pure sludge (kg/m3)
mass of soil in zone of incorporation (mixing zone) for sludge (Mg/ha)
mass of contaminant in soil at time / i(kg/ha)
molecular weight of contaminant (g/mol)
empirically derived exponent (diimensionless)
number of years sludge is applied to land (yr)
number of years sludge is applied to land used to grow crop / (yr)
concentration of pollutant/ in sludge Og/g) and
average rate of emissions *from the soil surface (kg/m2)
average rate of emissions from the soil surface in first second after
application of sludge (kg/m2-sec)
LF
LS
Mc.
Met
Mew
M:
M,
LF
•LS
MN
M,
Mp
M,
MCL
ME..
MS
MSH
M,
MW
n
N
Na
Na,
XXVIll
-------
List of Mathematical Symbols (continued)
Nay = average rate of emissions from the soil surface in first year after
application of sludge (kg/m2-yr)
NCI, = individual risk from non-carcinogenic pollutant J, expressed as the ratio
of exposure to Risk Reference Dose (percent)
NCP = "umber of persons exceeding RfD;of pollutant/ due to sludge exposure
j
number of persons exceeding RfD of pollutant/ without exposure to
sludge (i.e., with background intake alone) (persons)
NCPJ2 - number of persons exceeding RfD of pollutant/ with sludge exposure
and background intake combined (persons)
NR = net rechairge (m/yr)
°* = emission rate for organic pollutant/ for a unit sludge feed rate for
.facility of same type as facility p (g/sec per DMT/sec)
" — 4ndex-f0F teeiflerator-faeifities-
P' ~ percent solids in liquid layer of surface impoundment (kg/kg)
£2 ~ percent solids in sediment layer of surface impoundment (kg/kg)
lf ~ ratl° of contaminant concentration in fillet to whole fish (dimensionless)
p« ~ number of persons in cell/ exceeding RfD for pollutant/ (persons)
^ ~ percent liquid in the water column of surface water body (kg/kg or
dimensionless)
p1 = percent solids in the water column (kg/kg or dimensionless)
PA - total annual loading of contaminant to site (kg/yr)
PbB = concentration of lead in human blood ftig/d/)
PbB' = concentration of lead in human blood without exposure to lead from
sludge Otg/dO
PbB, = concentration of lead in human blood with exposure to lead from sludge
6
Pop = total exposied population (persons)
pPPi = population living in cell i (persons)
q.j = human cancer potency for pollutant j (mg/kg-day)'1
q = human cancer potency (mg/kg -day)'*
q"c = time-weighted average rafe of contaminant volatilization from a
monofill (g/m2-sec)
qco ~ rate of contaminant volatilization from a covered monofill cell
(g/m2-sec)
qun = rate of contaminant volatilization (emission rate) from an uncovered
monofill ceil (g/m2-sec)
^ = rate of inflow for sludge into a. surface impoundment (mVsec)
vo - rate of outflow from a surface impoundment (mVsec)
"«p = seepage rate for both liquid and sediment layers (m/sec)
r - distance from center of sludge disposal facility to the receptor's location
(m) , ' .
R = ideal gas constant (m3-atm/mol-K) •
xxix
-------
RAC
RCP
RCS
RE
RF
RFA
RL
RWC
S
S,
S2
se
sr
"wr
Sc,
SC
SMA
SRR
t
t.
tun
T
TF
TP
TSS
u
u,0
List of Mathematical Symbols (continued)
combined removal efficiency for pollutant/ for furnace and control p
expressed as fraction of original contaminant remaining in emissions
(dimensionless)
reference air concentration G«g/m3)
reference concentration of contaminant in seepage beneath a surface
impoundment (kg/ra3)
reference concentration of contaminant in surface-disposed sludge
(mg/kg) ;
relative effectiveness of exposure (dimensionless)
retardation factor (dimensionless) '
reference annual flux of contaminant beneath the site (kg/ha-yr)
Risk Reference Dose for pollutant/ (mg/kg-day)
risk level (incremental risk of cancer per lifetime)
reference water concentration (ing//)
intermediate variable used to calculate volatile emissions from soil
solids concentration in the liquid layer of a surface impoundment
(kg/m3)
solids concentration in the sediment layer of a surface impoundment
(kg/m3) ;
effective water saturation (dimensionless)
specific storage (m"')
sediment delivery ratio for the SMA (dimensionless)
water saturation (dimensionless)
residual water saturation (dimensionless)
sediment delivery ratio for the watershed (dimensionless)
Schmidt number on gas side (dimensionless)
Schmidt number on liquid side (dimensionless)
mass of sludge contained in one hectare of monofill (kg/ha)
sludge management area
source-receptor ratio (sec/m)
time (sec or yr)
duration of emissions (sec)
time that typical monofill cell contains sludge without soil cover (yr)
temperature (K)
active lifetime of surface disposal facility (sec)
duration of "square wave" for approximating the loading of
contaminant into the top of the unsaturated soil zone (yr)
total suspended solids content of the stream (mg//)
wind speed (m/sec) :
wind velocity at 10 meters altitude (m/sec)
rate of uptake of pollutant/ into tissue of crop i (mg/kg dry weight per
kg/ha) . :
XXX
-------
List of Mathematical Symbols (continued)
ujk = rate pf uptake of pollutant./ into meat or dairy product k per unit of
concentration in animal's food (mg/kg dry weight in animal tissue per
mg/kg dry weight in feed)
v = vertical teim for dispersion of contaminant in air (dimensionless)
Vd = darcy velocity (m/sec)
vh = regional velocity of horizontal groundwater flow (m/sec)
vi = superimposed radial velocity from water seeping from the impoundment
(m/sec)
vv = vertical velocity due to the source (m/sec)
vt = volume of air in soil (m3)
V, = volume of solids in soil (m3)
vi = total volume of soil (m3)
vv = volume of void space in soil (m3)
vw = volume of water in soil (m3)
x = distance from sludge management area.or disposal facility to receDtor
(km) , V
xy — lateral virtual distance to receptor location (m)
z = vertical coordinate in unsaturated zone (ra)
xxxi
-------
1. INTRODUCTION
«
The Environmental Protection Agency (U.S. EPA) is currently developing comprehensive
regulations to control the use and disposal of municipal sewage sludge under Section 405(d) of
the Clean Water Act (CWA). This, project supports that regulatory effort by providing estimates
of both current human health risks and the benefits of controlling these risks for the major sludge
management practices. The key sludge management practices examined are consistent with the
CWA 405(d) regulatory effort and include incineration, surface disposal, and land application
(food chain, non-food chain and distribution and marketing for residential uses).
In general, this report provides quantitative estimates of human health risks. We focus
on both carcinogenic and non-carcinogenic effects from exposure to sludge pollutants of concern
as identified by EPA's Office of Science and Technology (U,S. EPA, 1987d). We rely on fate,
transport and exposure methodologies developed by EPA's Office of Research and Development
and recommended by the Agency's scientists. Those compounds exhibiting carcinogenic effects
are examined using the best available dose-response estimates obtained from the Agency's
Integrated Risk Information System (IRIS). Non-carcinogenic substances are evaluated in terms
of the likelihood that exposure will exceed threshold valuesior Risk Reference Doses (RfDs).
There are several important considerations excluded from the quantitative risk estimates.
Most importantly, we did not attempt to a characterize ecological risks. Currently, available
methodologies and data do not permit an adequate analysis of such risks. Neither have we
attempted to estimate risks associate with exposure to pathogenic organisms. While the sewage
sludge regulations do address pathogens, estimating health risks from pathogens was not within
the scope of this analysis. This document also excludes the co-disposal of sludge with municipal
solid waste, because that practice will be controlled under Subtitle D of the Resource
Conservation and Recovery Act. In addition, indirect exposure routes, such as exposure due to
deposition from incineration of sewage sludge, were not assessed.
In this chapter we develop the general analytic framework used for the analysis, including
first a generic approach to risk assessment and risk-based benefits analysis and the application
of this generic framework to municipal sewage sludge. The remaining chapters of this report
provide detailed descriptions of the methodology and results for each of the major sewage sludge
use and disposal practices to be reflated under CWA 405(d).
1.1 GENERAL FRAMEWO1KK FOR RISK ASSESSMENT AND RISK-BASED
BENEFITS ANALYSIS
The aggregate risk assessment is intended to estimate both the expected national human
health risks associated with current use or disposal of sludge ("baseline" conditions) and the
benefits of the regulation measured in terms of estimated reductions in human health risks.
Methods for determining these risks differ for each of the management practices (incineration,
surface disposal and land application). In general, we use a sample of plants from the analytical
component of the National Sewage Sludge Survey (NSSS) to represent the larger universe of
1-1 .
-------
actual facilities. Estimated health risks from these plants under current practices were scaled
according to weighting factors from the survey to calculate risks at the national level.
*
In general, risk assessment couples information on hazard identification and chemical
specific assessments of dose-response relationships, with an exposure assessment based on fate
and transport estimates and health (mortality/morbidity) jand ecological effects. Benefits are
estimated as the reduction in baseline risk associated withiimposing regulatory controls. These
are estimated by re-computing risks based on implementation of the regulation and represent the
change (usually reduction if benefits are positive) in risks resulting from the relevant controls.
The remainder of this section describes the approach to human health risk assessment
used in this analysis. This process was outlined originally by the National Academy of Sciences
(NAS, 1983) and was established as final Risk Assessment Guidelines as published in the
Federal Register (U.S. EPA, 19iB6a). Five types of guidelines were issued:
Guidelines for Carcinogen Assessment,
Guidelines for Estimating Exposure,
Guidelines for Miiitagenicity Risk Assessment,
Guidelines for Health Effects of Suspect Developmental Toxicants, and
Guidelines for Heilth Risk Assessment of Chemical Mixtures.
The Risk Assessment Methodology consists of four distinct steps: hazard identification
dose-response evaluation, exposure evaluation, and finally, characterization of risks Each of
these steps is discussed below.
Hazard Identification
Hazard identification consists of gathering and evaluating all relevant data that help
determine whether a chemical poses a specific hazard, and making a qualitative evaluation based
on the type of health effect produced, the conditions of exposure and the metabolic processes
within the body that govern chemical behavior. Its goal is to determine whether it is
scientifically appropriate to infer that effects observed under one set of conditions (e.g., in
experimental animals) are likely to occur in other settings (e.g., in humans), and whether data
are adequate to support a quantitiitive risk assessment.
Information on the toxic properties of chemical substances is obtained principally from
animal studies and controlled epidemiological investigations' of exposed human populations. The
use of animal toxicity studies is baised on the longstanding assumption that effects in humans can
be inferred from effects in animals. The usual starting point for such investigations is the study
of acute toxicity in experimental animals. Acute toxicity studies are necessary to calculate doses
that would not be lethal to animals used in studies of longer duration, and typically involve a
single dose or exposure of very short duration (e.g., hours of inhalation). Acute toxicity is
usually expressed in terms of its LDJO, defined as the lethal dose on average for 50% of an
exposed population. Substances exhibiting a low LDM (e.g., for sodium cyanide, 6.4 mg/kg)
are more acutely toxic than those with higher values (e.g., for sodium chloride, 3000 mg/kg).
1-2
-------
In some experiments, animals are exposed continuously for several weeks or months
(sub-chronic toxicity studies) or up"to their full lifetimes (chronic toxicity studies). One of the
goals of these types of studies is to determine the "no observed adverse effect level" (NOAEL),
which is the dose at which ho effect is seen. In addition to this important objective, chronic
toxicity studies seek to identify specific organs or systems of the body that may be damaged by
exposure to a chemical and to identify the specific disease or abnormality that a chemical may
produce (e.g., cancer, neurotoxie effects.). :-
Animal studies often vary widely in design and implementation. Although standardized
tests have been developed for various types of toxicity (e.g., National Toxicology Program
carcinogenic bioassays), some teists are conducted using specialized study designs for situations
where established guidelines have not been developed. Factors that need to be considered when
evaluating the design and results of animal toxicity studies include selection of animal species
(e.g., rats, monkeys), dose and duration (e.g., acute vs. chronic), and route of exposure (e.g.,
ingestion, inhalation).
Epidemiological studies are useful for determining whether a chemical poses a hazard to
human health. These studies compare the health status oJF a group of persons who have been
exposed to a suspected causal agent with that of a comparable non-exposed group. Most are
either case-control studies or cohort studies. In case-control studies, a group of individuals with
specific disease is identified (cas<5s) and compared to individuals without the disease (controls)
in an attempt to ascertain commonalities in exposures they may have experienced in the past.
Cohort studies start with a study population, or cohort, considered free of the disease under
investigation. The health status of the cohort known to have a common exposure is examined
over time to determine whether any specific condition or cause of death is indicated which is
greater than what might be expected due to other causes.
Because epidemiological studies assess the effects of a chemical on a human population
directly, they are considered to elicit more convincing evidence of risks to human health than
the results of animal studies. In general, a robust epidemiological study requires:
• determination of ejcposure levels, including the degree and duration of exposure,
« designation of control groups that control for other risk related factors that affect
exposure and/or health status,
• availability of systematic information on health effects,
« duration of investigation sufficiently long to detect certain health effects, such as
cancer, and
• adequate sizes of studied populations to provide statistical power of detection.
The next step in hazard identification is to combine the pertinent data to ascertain the
degree of hazard associated with each chemical. In general, EPA uses different approaches for
the qualitative assessment of risk or hazard associated with carcinogenic versus non-carcinogenic
effects. i
1-3
-------
EPA's guidelines for carcinogenic risk assessment (U.S. EPA, 1986a) group all human
and animal data reviewed into the following categories based on degree of evidence of
carcinogenicity: . •
• Sufficient Evidence,
• Limited Evidence (e.g., in animals, an increased incidence of benign tumors
only),
• Inadequate Evidence, ;
• No Data Available, and
• No Evidence of Carcinogenicity.
Human and animal evidence of carcinogenicity in'these categories is combined into a
weight-of-evidence classification scheme (described in Table 1-1) that includes the following
groups:
• Group A - Human carcinogen,
• Group B - Prolbable human carcinogen,
- Bl ffigher degree of evidence,
- B2 Lower degree of evidence,
• Group C - Possible human carcinogen,
• Group D - Not classifiable as to human carcinogenicity, and
• Group E - Evidence of non-carcinogenicity.
Table 1-1
Classification Scheme for Weight of Evidence for Carcinogenicity
Human
Evidence
Sufficient
Limited
Inadequate
No Data
Evidence of
No Effect
Animal Evidence
Sufficient
A
Bl
B2
B2
B2
Limited
A
Bl
C
C
C
Inadequate
A
Bl
D
D
D
No Data
A
Bl
P
D
D
Evidence of
No Effect
A
Bl
D
E
E
Group B (probable human carcinogens) are usually divided into two subgroups in which
Bl is used to categorize the chemicals for which there is some limited evidence of
1-4 «
-------
carcinogenicity from epiderniological studies, and B2 for which there is sufficient evidence from
animal studies but inadequate evidence from epidemiological studies. EPA treats chemicals
classified in categories A and B as suitable for quantitative risk assessment. Chemicals classified
as Category C received varying treatment with respect to the dose-response assessment and are
determined on a case-by-case biisis. Chemicals in Groups D and E do not have sufficient
evidence to support a quantitative dose-response assessment.
The following factors are evaluated based on a judgment of the relevance of the data for
a particular chemical: :
quality of data,
resolving power of the studies (significance! of the studies as a function .of the
number of animals or subjects),
relevance of route and timing of exposure,
appropriateness of dose selection,
replication of effects,
number of species examined, and
availability of human epidemiologic study data.
As with weight-of-evidence for carcinogenicity, this information is not used to
quantitatively estimate risk. Rather, hazard identification characterizes the body of scientific
data in such a way as to provide both a determination as to whether a chemical is a hazard and
if quantitative assessment is appropriate. For non-carcinogenic health effects, the Agency's
weight of evidence categories have not been formalized. '
Dose-Response Evaluation
The next step in the risk assessment methodology is to estimate the dose-response
relationships for the chemical under review. The evaluation of dose-response data involves
quantitatively characterizing the connection between exposure to a chemical (measured as
quantity and duration) and the extent of toxic injury or disease. In most cases, dose-response
relationships are estimated based on animal studies because even good epidemiological studies
rarely have reliable information on exposure.1 Therefore, this discussion focuses primarily on
dose-response evaluations based on animal data.
There are two general approaches to dose-response evaluation, depending on whether the
health effects are based on threshold or non-threshold characteristics of the chemical. In this
context, thresholds refer to exposure levels below which no adverse health effects are assumed
to occur. For effects that involve the alteration of genetic material (including carcinogenicity
and mutagenicity), the Agency's ]position is that effects may take place at very low doses, and
therefore are modeled with no thresholds. For most other biological effects, it is usually (but
not always) assumed that "threshold" levels exist.
'An important exception is lead, which will be discussed in detail in Chapter 2.
1-5
-------
For non-threshold effects, the key assumption is that the dose-response curve achieves
zero risk only at zero dose. A mathematical model is used to extrapolate response data from
doses in the observed (experimental) range to response estimates in the low dose ranges.
Scientists have developed several mathematical models to estimate low dose risks from high dose
experimental results. Each model is based on general theories of carcinogenesis rather than on
data for specific chemicals. Hie choice of extrapolation model can have a significant impact on
the dose response estimate. The Agency's cancer assessment guidelines recommend the use of
the multistage model, which yields estimates of risk that are conservative, representing a
plausible upper limit of risk. With this approach, the estimate of risk is not likely to be lower
than the true risk (U.S. EPA 19!56a). . i •
EPA has acknowledged tliat a procedure does not yet exist for making the "most likely"
or "best" estimates of risk within the range of uncertainty defined by the upper and lower limits.
It can be shown that using the maximum likelihood estimates of dose-response curves may be
quite different than those product by the upper confidence limit. For example, the maximum
likelihood estimate for formaldehyde yielded dose-response curve estimates five orders of
magnitude less stringent than those produced by the upper confidence limit estimate, yet both
were derived with the multistage model.
The Agency also provides guidance on extrapolating dosages from animal species to
humans. Several methods have been proposed including:
• milligram per kilogram body weight per day,
• milligram per squaire meter body surface area per day, and
• parts per million in the air, water or diet.
The Agency advocates the use of the surface area approach, the most scientifically conservative,
for extrapolating results from animals to humans (U.S. EPA, 1986a).
The resulting cancer potency estimate, whiich is referred to by the Carcinogenic
Assessment Group as q*, is the quantitative expression derived from the linearized multi-stage
model and represents a plausible tapper- bound estimate to jthe slope of the dose-response curve
in the low dose range. The q* is expressed in terms of risk-per-dose, and has units of
(mg/kg-day)'1. These values should be used only in dose ranges for which the statistical dose-
response extrapolation is appropriate.
Dose-response relationships are assumed to exhibit threshold effects for systemic toxicants
or other compounds exhibiting non-carcinogenic, non-mutagenic health effects. Dose-response
evaluations for substances exhibiting threshold responses involve calculating what is known as
the Risk Reference Dose or Reference Concentration (RfD/RfC). This measure is used as a
threshold level for critical non-camcer effects below which a significant risk of adverse effects
is not expected.
The RfD/RfC methodology involves taking the experimental dose at which little or no
effect is observed and dividing by an appropriate uncertainty factor. These experimental levels
can consist of the No Observed Effect Level (NOEL), No Observed Adverse Effect Level
1-6
-------
(NOAEL), Lowest Observed Effect Level (LOEL) or lowest Observed Adverse Effect Level
(LOAEL) and can be derived from either laboratory animals and/or human epidemiological
studies. The safety factory ranging from 100 to 10,000; are used to extrapolate from acute to
chronic effects, inter-species sensitivity, variation in sensitivity in human populations and
extrapolation from a LOAEL to a NOAEL. Their magnitude can vary from 100 to 10 000
depending on the nature and quality of the data from which the experimental levels are derived'
Ideally, for all threshold effects!, a set of route-specific arid effect-specific thresholds should be
developed. If information is available for only one rouie of exposure, this value is used in a
route-to-route extrapolation to estimate the appropriate threshold. Once these values are derived
the next step is to estimate actual human (or animal) exposure. '
Exposure Evaluation '
Exposure evaluation involves the nature and size of the population exposed to a substance
and the level, timing and duration of their exposure. The major areas to be evaluated when
estimating exposures are: '
source assessment, •
pathways and fate analysis,
estimation of environmental concentrations,
population analysis, and <
integrated exposure analysis.
Ideally, the source assessment should account for all mass flows of a chemical from creation to
destruction. Environmental releases should account for temporal (e.g., seasonal) and geographic
distributions in all environmental media. For this analysis, the source assessment for pollutants
in sewage sludge begins with the characterization of pollutant concentrations, the quantities of
sludge generated, and sites for the use or disposal of sludge.
A pathway and fate analysis describes how a chemical may be transported from a source
to the potentially exposed population. This part of exposure assessment involves both an
analysis of chemical transport aiad transformation, and an identification of principal pathways
of exposure. Transport refers to physical properties, such as volatilization, dispersion or
advection, that may effect the chemical's ultimate fate. Transformation accounts for chemical
processes such as hydrolysis, photolysis or biodegradation that cause parent compounds to break
down into progeny compounds. This part of exposure assessment should also account for factors
such as inter-medk transfers and should identify those pathways that may result in the greatest
potential for exposure.
Two distinct approaches jire available for estimating environmental concentrations:
• direct measurement of levels of chemicals (monitoring), and
• use of mathematical models to predict concentrations.
Measurements are a direct and preferred source of information for exposure analysis. However,
they are usually expensive and are often limited geographically. The best use of such data is
1-7
-------
to calibrate mathematical models that can be more ! widely applied. When estimating
concentrations using mathematical models, the analyst needs not only to account for physical and
chemical properties but also to document mathematical properties (e.g., analytical integration
vs. statistical approach), spatial properties (e.g., one-, two- or three dimensions) and time
properties (steady-state vs. notisteady-state). Also, to the extent possible, selection of the
appropriate fate and transport model should follow guidelines specified for particular media.
For example, the Guidelines on Air Quality Models (U.S. EPA, 1986b) provide guidance for
selecting models for the dispersion of pollutants in ambient air.
Population analysis involves a description of the size and characteristics (e.g., age or sex
distribution), location, and habits (e.g., food consumption or workplace) of potentially exposed
human and non-human populations. Census and other survey data are useful in identifying and
describing human populations ejcposed to a chemical.
Integrated exposure analysis involves the calculation of exposure levels along with a
description of the exposed populations. An integrated exposure analysis quantifies the contact
of an exposed population to each chemical under investigation via all routes of exposure and all
pathways from the sources to the exposed individuals. Finally, uncertainty should be described
and quantified to the extent possible.
For risk assessments involving chronic exposure, human exposure is calculated as:
ADLE = 9
365xLSxBW
where:
ADLE = average daily lifetime exposure (mg/kg-day),
D = total dose (mg),
LS = average lifuspan (yr), and
BW = body weight (kg).
The total dose can be expanded as:
D - C x CR x ED x AF
where:
C = Contaminant Concentration, or the concentration of the chemical in the
medium (air, food, drinking water) contacting the body. Typical units are
mass/volume (e.g., jig/1 or /*g/m3) or mass/mass (e.g., mg/kg).
CR = Contact Rate, or the rate at which thb medium contacts the body (through
inhalation, ingestion or dermal contact). Typical units are mass/time
(e.g., mg/dlay) or volume/time (e.g., m3 or I/day).
ED = Exposure Duration, or the length of time for contact with the chemical
(e.g., lifetime).
1-8
-------
AF = Absorption FractionT or the effective portion of total chemical contacting
and entering (he body. Entering the b6dy means that the chemical crosses
one of the three exchange membranes: alveolar membrane,
gastrointestinal tract, or skin.
Characterization of Risks
The final step in the risk assessment methodology is risk characterization. This step
essentially involves combining the information developed in hazard identification, dose-response
assessment and exposure assessment to derive quantitative estimates of risk. Qualitative
information should also accompany the numerical risk estimates, including a discussion of
uncertainties, limitations, and assumptions. It is useful to distinguish methods used for
chemicals exhibiting threshold effects (i.e., most non-carcinogens) from those believed to lack
a response threshold (i.e., carcinogens).
EPA has recently developed Guidance on Risk Characterization for Risk Managers and
Risk Assessors (U.S. EPA, 1992b) that defines three types of descriptors of risk that should be
developed as part of a risk assessment: (1) individual risk, which should include both the central
tendency and high end portions of the risk distribution, (2) important subgroups, such as highly
exposed or-highly-suseeptible-groups or-individuals, and (3) population-risks. EPA specified
recently that the definition of "high end" should be used to describe a plausible estimate
associated with a value above 90th percentile of the actual distribution. Thus, the highly
exposed individual (HEp is differentiated from worst case, Maximally Exposed Individual (MET)
or bounding estimates in that it should not overestimate risk's to the actual population.
For carcinogens, individual risks generally are represented as the probability that an
individual will contract cancer in a lifetime as a result of exposure to a particular chemical or
group of chemicals. Population risks are usually estimated based on expected or average
exposure scenarios (unless information on distributions of exposure is available.) The number
of persons above a certain risk level, such as 10* (one in a million), or a series of risk levels
(IQr5,104, etc.) is another useful descriptor of population risks. Thus, individual risks also may
be presented using cumulative frequency distributions where the total number of persons
exceeding a given risk level is plotted against the individual;risk level.
i •
For non-carcinogens, dose-response data above the threshold are usually lacking.
Therefore, risks are characterized by a comparison to the threshold level by the ratio of dose or
concentration to the threshold level. Aggregate population risks for non-carcinogens can be
characterized by the number of persons exposed above the RfD or RfC. The same approach can
be used to assess both acute and chronic hazards. For assessing acute effects, the toxicity data
and exposure assessment methods must account for the appropriate duration of exposure.
For carcinogens, risk to an individual can be represented as the maximum probability that
an individual will develop cancer in a lifetime as a result of exposure to a particular chemical
or group of chemicals. This probability is calculated from the estimated dose (or concentration):
where:
1-9
-------
cij = q; EXP.
*
Cij = incremental cancer risk from contaminant J for exposed individual
(incremental risk of developing cancer per lifetime of exposure),
q = human cancer potency for pollutant/' (mg/kg-day)-1, and
GXP, = exposure to contaminant/' (mg/kg-day).
Dose is a function of the concentration in environmental media, the contact rate for the chemical
(inhalation rate, dermal contact rate, food consumption rate, etc.) and its rate of biological
absorption.
Individual risk from a non-carcinogen is expressed as the ratio of the dose to the
Reference Dose (expressed as a percentage):
EXP,
NCI, = i x 10096
ftfDj
where:
Nai - ^dividualrisk from non-carcinogenic^fliitant/', expressed as the ratio of
exposure to Risk Reference Dose (percent), and
RfDj = Risk Reference Dose or reference concentration of pollutant/'.
The pattern of environmental contamination from a given chemical varies across
geographic areas. As a result, different portions of the population are exposed at different
levels. The population risk assessment for both carcinogens and non-carcinogens involves
calculating the risk estimates for twery combination of population and concentration or dose2
These risk estimates are then summed across the entire area of concern. For a particular
exposure group, the aggregate risk for a carcinogen is calculated from:
CP = ^4 EXP POP
where:
CP = expected number of incremental cancer cases for this exposure group
(cases/yr),
LS = average human lifespan (yr), .-,--.- ,
EXP = exposure for persons in group (mg/kg-day), and
POP = population of exposure group (peirsons).
2The expected number of cani^r cases is computed per year so risk values based on a
lifetime exposure must be adjusted by the average duration of a human lifetime.
1-10
-------
These estimates are summed across all exposure groups for an estimate of aggregate cancer
risks. :
The aggregate risk for a non-carcinogen is defined as the total number of persons
exposed to a yearly average concentration greater than the reference concentration (or dose).
For each exposure group, the dose is compared to the reference dose. Aggregate risk is
calculated by summing the number of persons whose dose: exceeds the RfD.
Information on the distribution of risk is often also useful to provide a means of
combining information on individual risk levels with the expected size of populations exposed
at those levels. To accomplish this, random simulation or other methods can be used to generate
a distribution of risk based on the individual distributions of the individual input parameters.
This approach provides a means of incorporating uncertainty directly into the estimates.
1.2 APPLICATION OF METHODOLOGY TO THE USE AND DISPOSAL OF
SEWAGE SLUDGE
The objective of our analysis is to estimate the reduction in human health risks due to
the regulatory standards and controls for the use and disposal of sewage sludge. The first step
in this analysis was to estimate baseline human health risks. Once baseline risks were defined,
the appropriate control options were identified and the expected changes in risks were quantified!
This information can also be coupled with cost information to produce cost-risk measures of the
regulation.
The key inputs for estimating baseline risks include information on the sources (publicly-
owned treatment works, or POTWs), sludge pollutants and ultimate disposal site characteristics.
Baseline risks from the major sludge disposal options were characterized based on initial sludge
concentrations, the actual quantity of sludge generated for each POTW and a number of different
environments, depending upon disposal option. All these components are discussed in more
detail later in this chapter.
On the basis of available information, we modeled the fate and transport of the key
pollutants for the primary pathways of human health exposure. We then estimated the potential
populations exposed, coupled this information with chemical-specific dose-response data to
characterize baseline human health risks to both the highly exposed individual and to the
aggregate population as a whole. This information is presented by pollutant, exposure pathway
and disposal practice.
Once baseline risks were derived, we adjusted our assumptions for management practices
to correspond with responses to regulatory controls., We then repeated all calculations to
estimate the change in risk as a result of regulatory controls. This change (expressed as the
number of cases avoided per year) provides a benefit measure for the regulation.
The remainder of this section briefly discusses key components of the risk assessment
methodology, including:
1-11 • , '
-------
sludge pollutants, :
disposal options," !
fate and transport modelling, : .
exposure pathways, :
exposed populations,
health effects, • j
risk characterization, ;
compliance strategies, and i
benefit analysis.
1.2.1 Sludge Pollutants
Municipal sewage sludge typically contains a wide array of pollutants, from heavy metals
to organic pollutants and pathogenic organisms. A1982 EPA study identified over 200 chemical
constituents from a number of sludge samples and studies (Booz-Allen & Hamilton, 1982).
Based on that list, EPA conducted a preliminary screening analysis in 1983 to select a subset of
pollutants of concern based on the frequency of occurrence of these constituents in sludges (from
EPA's 40-City study) and on available human health and ecological toxicity infonnation. Expert
panels were then convened for each of the major disposal options to obtain consensus on whether
the Agency selected the correct list of pollutants or whether additions and/or deletions should
be made. The collective judgments of these panels resulted in a list of 50 pollutants that were
deemed to be of concern.
During 1984 and 1985, EPA's Office of Water developed a series of profile documents
for each of the 50 pollutants. For each of the key disposal practices, these profile documents
examined, using screening-level fate, transport and exposure models and generally worst case
assumptions, the propensity of any of the pollutants to pose a hazard. Human health hazard was
determined for carcinogens by using an equivalent IxlO* cancer risk determined to be de
minimus, and for Eon-carcinogens using available RfD information. Ecological effects were also
examined using available threshold effect criteria, such as water quality criteria. The objective
of this exercise was to narrow further the list of pollutants of possible concern.
This effort yielded the revised list of pollutants shbwn in Table 1-2. Note that in this
report, as in the proposed sludge regulations, not all practices are considered for each of the
sludge pollutants of concern. Tills is because the sludge profile screening analysis documents
were able to eliminate certain disposal practices from further consideration. Again, this
screening analysis was based on a worst-case exposure and risk assessment for each of the
pollutants listed. The pollutants
-------
Table 1-2
Sludge Pollutants by Management Practice
Aldrin/Dieldrin
Arsenic ' " '
Benzene
Benzidine
Benzo(a)pyrene
Beryllium
Bis(2-ethylhexyl)phthalate
Cadmium
Carbon Tetrachloride
Chordane
Chloroform
Chromium
Copper
DDT/DDD/DDE
Fluorine
Heptachlor
Hexachlorobenzene
Hexachlorobutadiene
Iron .
Lead
Lindane
Mercury
Molybdenum
Nickel
Nitrosodimethylamine
PCBs
Selenium
Toxaphene
Tricbloroethylene
Vinyl Chloride
Zinc
Incineration
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
. Land Application Surface Disposal
! x
X
x
1 X
'; X
X
X
; X
X
X
X
x
X
X
X
; .... 3_._...^..._..
... 1. _. x ., - •.
,X _ ,....,.
: x '
X
'
X
X
X
X
X
X
X
.... ...
X
X
X
X
X
X
X
X
X
X
X
"Organic pollutants considered for inciniwrators include additional products, of incomplete combustion. A list of
pollutants considered for this analysis is provided in Table 3-9 of Chapter 3.
1-13
-------
1.2.2 Disposal Options '
The current 405(d)'sludge regulation addresses all use or disposal practices for sewage
sludge, with the following exceptions:
• landfilled sludge that is co-disposed widti- municipal solid waste (this will be
addressed by the RCRA Subtitle D regulation),
« ocean disposal (this practice is being phased out completely), and
• co-incineration of sludge with municipal ;solid waste (this will be addressed at
a later date).
In the regulation and in this report, use and disposal practices are grouped into three
general categories:
* incineration,
• land application (including food chain, non-food chain, and residential uses)
and '
• surface disposal..
To determine the quality iind quantity of sludge managed with each of these options, and
to address serious shortcomings in data used for previous analyses of costs and benefits from
regulating sewage sludge, the .EPA conducted the National Sewage Sludge Survey. This survey
conducted between August, 1988! and September, 1989, characterized pollutant concentrations
and sewage sludge practices, and consisted of two major- components: a questionnaire which
covers management practices arid related data, and an analytical survey which focusses on
pollutant concentrations found in isewage sludge samples. Candidate POTWs for the survey were
based on over 11,000 POTWs in the U.S. and Puerto Rico, which were identified in the 1986
Needs Survey as having secondary or advanced treatment. A random sample was then stratified
to be representative based on estimated wastewater flow and sewage sludge disposal practices.
Out of the 479 POTWs randomly chosen to receive the questionnaire, 208 facilities were then
selected for the analytical survey. Survey weights were utilized to estimate national risks
associated with each practice.
The questionnaire component of NSSS was used to ;obtain general information about the
POTW to derive national estimates of the total quantities of sewage sludge generated and
estimates of treatment practices, siewage sludge use and disposal practices, quantities associated
with each practice, and sewage sludge treatment and disposal costs. The analytical component
of NSSS was designed to obtain sewage sludge samples for 419 pollutants, which were selected
from several existing EPA regulations, including the Clean Water Act Section 307(a) priority
pollutants, Appendix Vn pollutants for the Resource Conservation and Recovery Act (40 CFR
Part 264), toxic compounds identified in the Domestic Sewage Study, and other contaminants
of potential concern for municipal sewage sludge. Table 1-3 depicts the current disposal
practices by number of POTW and by percent of total sludge generated.
1-14
-------
Table 1-3
Quantity of Sludge by Management Practice
INCINERATION
LAND APPLICATION
Food Chain Agriculture
Residential
Other
Total
SURFACE DISPOSAL
Monofdls
Other
Dedicated
Total
OTHER0
TOTAL
Quantity of Sludge
(DMT/yr)*
820,000
970,000
130,000
350,000
1,500,000
,
140,000
88,000
240,000
470,000
1,500,000
4,200,000
Percent of Totalb
20%
23%
3%
8%
34%
3%
.2*
6%
11%
35%
100%
•Quantities reported in two significant figures.
bDue to independent rounding, the percentages may not sum to totals.
""Other" includes not regulated (31% of total), ineligible out-of-business, or unknown.
1-15
-------
1.2.3 Fate and Transport Modeling
•
This study couples fate and transport models to predict risk for each sewage sludge use
disposal practice (incineration, bind application and surface disposal). The models we use to
predict baseline risks and benefits are consistent with those used by the Office of Science and
Technology to derive numerical criteria. These models, together with accompanying
assumptions, have been-peer reviewed both within the Agency and externally to EPA.
Table 1-4 summarizes the models used to predict fate and transport of sludge constituents
by disposal practice and medium of concern. These models combine data for pollutant-specific
physical/chemical properties (e.g., decay rate, solubffity, etc.) with data for the characteristics
of use or disposal sites to predict chemical concentrations at selected point(s) in time and space.
Therefore, the combination of pollutant concentrations in sludge, amount of sludge generated,
the sludge disposal option and the region or site where disposal is practiced, can all greatly
affect fate and transport model predictions. Moreover, while models chosen are among the best
available, they often do not consider all relevant factors, or not all scientific and engineering
factors are available to characterize the number and range of disposal scenarios and
environments possible across thesie United States. Only after many of these models are field-
tested and validated with monitoring data from sludge disposal, can we hope to narrow the
uncertainty in these predictions. More detailed explanations concerning each model and
application are included in later chapters of this report.
1.2.4 Exposure Pathways
Once pollutant concentrations are predicted using the fate and transport models, we
determine the primary routes of human exposure for each of the disposal options. Consumption
of drinking water; inhalation of airborne pollutants; ingestion of grain, vegetables, meat, poultry,
dairy products, and fish; and direct ingestion of treated soil are the main routes of exposure from
the use or disposal of sludge.
Exposure through drinking water can occur from the contamination of either surface
water or groundwater supplies. Pollutants from both land-applied and surface-disposed sludge
can migrate to aquifers beneath the use or disposal site, where they can be transported laterally
to nearby drinking water wells. For land application, soil eroding from the application site can
cany adsorbed pollutant to nearby surface'water bodies used as sources for drinking water.
Exposure through inhalation can occur when pollutants from an incinerator are released
through emissions and transported by wind to nearby human populations. Similarly, pollutant
volatilizing from land application or surface disposal sites can be transported to humans living
near the sites where sludge is used or disposed.
Exposure is also possible through dietary consumption of grains, vegetables, or other
products grown on sludge-amended soil. Less directly, pasture or grains from treated land can
be used to feed animals, which may produce meat, eggs or dairy products for human
consumption. Fish caught for consumption from surface water bodies polluted by eroded soil
1-16
-------
3
a
13
I
ggf
£ « Q £
P o c? o
-------
from land-application site may provide an additional source of potential dietary exposure
Finally, children can be exposed to pollutants from sludge if they directly ingest treated soil
An additional route of possible dietary exposure, the deposition of pollutants emitted by
incinerators and subsequent uptiike into crops or fish (or direct ingestion by children) is
excluded from this analysis.
1.2.5 Populations Exposed
Table 1-5 presents a summary by disposal option of the critical populations exposed to
sludge pollutants, and the data soim-ces used to estimate the sizes of those populations. For the
dietary pathways of exposure from land application, the entire U.S. population is assumed to be
exposed. Because of the complex distribution networks for grains, meat, and dairy products in
the U.S., it is virtually impossible to match individual land application sites with particular
populations of consumers. This smalysis therefore assumes that all crops and animal products
produced with sludge are distributed uniformly in the national food supply. For residential uses
of land-applied sludge, we base our analysis on members of households using sewage sludge for
vegetable or ornamental gardening.
For incineration, those persons living nearest to sludge incinerators are assumed to be
most vulnerable. As will be explained in more detail in Chapter 3, our analysis maps predicted
ground-level concentrations of emitted pollutants for about 800 locations within about 40 km of
each incinerator. Persons living within this range of at least one incinerator are considered the
exposed population. Similarly, for the groundwater and volatilization pathways, those
individuals living within about 3 km of a surface disposal or land application site are considered
for this analysis. Details are provided in Chapters 5 and 7;
1.2.6 Health Effects
For the sludge pollutants of concern, we attempt to identify the potential adverse health
effects that may be associated with chronic exposure. We rely primarily on findings from the
EPA's Office of Research and Etevelopment to estimate the effects of each pollutant. An
important distinction is the difference in methods used for carcinogenic and non-carcinogenic
compounds.
As discussed previously, for known and suspected carcinogens we rely on estimates of
potency from the EPA's Office of Health and Environmental Assessment (OHEA). OHEA uses
a linearized multistage procedure with zero as the threshold jto derive the plausible upper bound
slope (or #*) on the low dose portion of the dose-response curve. Table 1-6 and Table 1-7 list
estimates of human cancer potency for the organic pollutants and metals (respectively)
considered for this analysis. The toibles also include Risik Reference Doses for non-cancer health
effects. For non-carcinogens, we rely on the Agency's Risk Reference Dose (RfD) as the
threshold value obtained from EPA's Integrated Risk Information System (IRIS).
1-18
-------
Table 1-5
I
Exposed Populations by Management Practices
Population Exposed
Data Source
INCINERATION
Populations Residing Near
Sludge Incinerators
LAND APPLICATION
Food Chain Agriculture Entire U.S. Population
U.S. Census (1980) from
MARF (PC-GEMS)
Statistical Abstracts (1991)
All Non-residential
Residential Uses
Population Exposed through National Well-Water
Air, Groundwater, Surface Association Data Base
Water, or Fish ' (Wellfax)
Home Gardeners
National Home Gardening
Survey (1987)
SURFACE DISPOSAL
Population Exposed through National Well-Water
Air or Groundwater Association Data Base
(Wellfax)
1-19
-------
Table 1-6
Health Effects Data for Organic Contaminants in Sewage Sludge*
Risk
Reference Human Cancer
Dose Potency
(mg/kg-day) (mg/kg-day)'1
Aldrin/dieldrin
Benzene
Benzidine
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Carbon tetrachloride
Chlordane
Chloroform
DDT
Dimethylnitrosamine
Heptachlor
Hexachlorobenzene
Hexachlorobutadiene
Lindane
PCBs
Toxaphene
Trichloroethylene
Vinyl chloride
3xlO-s
NA
0.003
NA
0.02
0.0007
6xlO-5
0.01
0.0005
NA
0.0005
0.0008
0.002
0.0003
NA
NA
NA
NA
17
0.029
230
-7 *
t .J
0.014
0.13
1.3
0.0061
0.34
51
4.5
1.6
0.078
1.3
7.7
1.1
0.011
1.9!
* Note that human cancer potency values are for oral exposure.
exposure are shown in Table 3-10, Chapter 3.
HEAST = Health Effects Assessment Summary Tables, March
IRIS = Integrated Risk Information System, 1992
Level of
Evidence Reference
B-2
A
A
B-2
B-2
B-2
B-2
B-2
B-2
B-2
B-2
B-2
C
B-2
B-2
B-2
B-2
A
Potencies
1992
IRIS, IRIS
IRIS
IRIS, IRIS
IRIS
IRIS, IRIS
IRIS, IRIS
IRIS, IRIS
IRIS, IRIS
IRIS, IRIS
IRIS
IRIS, IRIS
mis, mis
IRIS, IRIS
IRIS,
HEAST
IRIS
IRIS
IRIS
HEAST
for inhalation
1-20
-------
Table 1-7
Health laffects Data for Metals in Sewage Sludge1
Risk Reference
Dose
(mg/kg-
-------
Consistent with EPA's final risk assessment guideline for Chemical Mixtures (U.S. EPA
1986a), we will assume additivity of cancer potency and risk assessment estimates across
pollutants for each disposal* option. We do not consider synergistic or antagonistic interactions
between pollutants.
One final point concerns the human absorption or the relative effectiveness (RE) of
exposure to each pollutant. RE is a unifless factor that indicates .the relative toxicological
effectiveness of an exposure by, for example, ingestion of drinking water. The value of RE may
reflect an observed or estimated difference in absorption rates between the inhalation and
ingestion routes, that is then assumed to translate into a difference in the toxicant's effectiveness.
When such information is not available, we assume an RE equal to one.
1.2.7 Risk Characterization
Once the results of our assessments are complete, there are two forms in which risks can
be presented: quantitative and qualitative. In depicting risks quantitatively, we present risk to
the average exposed individual (AEI), highly exposed individual (BET) and aggregate risks to
the population as a whole. In defining the HEI, we use reasonable worst-case assumptions
consistent with those used by the agency to derive numerical criteria. For risks to the AEI and
the total population, we generally rely on central tendency estimates for input parameters.
Specific choices are discussed in Chapters 3-7 of this report.
For estimating risks to the highly exposed individual, the calculation is straightforward:
ci. = EXPJ q;
where:
CIj = lifetime incremental risk of cancer from exposure to pollutant j from
sludge,
EXT, = incremental exposure to poEutanty for a particular population, in this case
the the highly exposed individual (mg/kg-day), and
qj* = human cancer potency for pollutant./' (mg/kg-day)'1.
This risk is usually expressed as the incremental probability that an individual will contract a
disease (usually cancer) over a lifetime of exposure. For non-carcinogens, we compare the
estimated dose to the RfD. Results are reported by listing exposure to each pollutant as a
fraction of the risk reference dose.
Individual risk to the AEI is estimated with identical methods, except the calculation is
based on average exposure to the population through a particular management practice or
exposure pathway. Exposure for the average individual (EXPj) is multiplied by the cancer
potency to estimate the average individual lifetime risk of developing cancer.
For aggregate population risk, the calculation is similar except that risks are calculated
separately for each individual sub-population and summed to derive an aggregate total. In
1-22
-------
addition, the size of the exposed population and the average human life expectancy is inserted
into the equation, so that the result is reported in units of 'incremental cases of cancer expected
per year: \ •
EXP, q," POP
CP. = J-*L
' LS
where:
CPj = expected incremental cases of cancer per year as result of exposure to
pollutant y from sludge (cases/yr),
EXPj = incremental exposure to pollutant j for particular population, in this case,
to a particular sub-population for which exposure has been estimated
(mg/kg-day), .:
POP = size of population with this level of incremental exposure to pollutant j
from sludge (persons), and
LS = average human lifespan (yr). '•
Results are summed across all populations for which exposure has been estimated to derive an
estimate of aggregate risk (in cases/yr). For non-carcinogenic pollutants, results are reported
by listing average exposure to each pollutant as a fraction of the risk reference dose, and by
estimating the number of persons for which exposure exceeds the risk reference dose because
of exposure to pollutants from sludge.
For any single management practice or exposure pathway, aggregate cancer risks (in
cases/yr) can be related to individual cancer risks for :the average exposed individual by
multiplying the individual risk by the size of the exposed population and dividing by the average
life expectancy. Conversely, multiplying the estimated! aggregate risk by the average life
expectancy and dividing by the size of the exposed populations gives the risk to the AEL This
same relationship holds for groups of management practices or exposure pathways (e.g., all
exposure pathways for land application, or all management practices for sludge). Where such
groupings are considered, the relevant population size is defined as the size of the population
of individuals exposed through one or more of the pathways under consideration.
1.2.8 Benefits Analysis and Compliance Strategies
Once the baseline risks and regulatory controls are identified, the final step involves the
quantification of the incremental change in risk as a result of the regulation. The estimate of
benefit we choose is the reduction of human health morbidity and/or mortality as a consequence
the regulatory requirement. This approach is best illustrated with a hypothetical example.
Suppose a hypothetical sludge incinerator emits a single pollutant./ with a cancer potency
of 0.07 (mg/kg-day)'1. We use a mathematical model to estimate expected concentrations of
pollutant j at locations near the incinerator and to map those concentrations with actual human
populations. For an exposed population of 100,000 residing near the incinerator, we estimate
that exposure averages 0.04 mg/kg/day, so baseline aggregate risk is calculated as:
1-23
-------
CP -
LS
= QM(mglkg-day) 0.07 (mgfkg-dayy1 100,000
70 ~
= 4(cases/yr)
Now assume that in response to Relation the incinerator bstalls additional emission controls
and that emissions of pollutant; are reduced 75 percent. Using the same mathematical model
we re-calculate exposure to find that it has been reduced to an average of about 0.01 mg/kg-dav'
so that the new estimate or aggregate risk is: : **sr™y,
j q* POP
LS
= 0.010»g/*g-
-------
As will be explained in Chapters 4-7, relatively few plants are expected to reduce their
use of land application or surfaces disposal as a result of the regulation. We have not attempted
to quantity potential benefits from refinements to management practices or possible pre-
treatment, except to suggest-tiiat these benefits are unlikely to exceed our estimates of baseline
risks for these practices.
1.2.9 Limitations/Uncertainty
Because this analysis utilizes data provided by the National Sewage Sludge Survey,
assumptions about the quantity and quality of sludge are believed to be more reliable than
assumptions used for previous analyses. Nevertheless, if the random, stratified sample of
facilities used for the analytic component of the NSSS (the basis for this analysis) is not truly
representative (in all respects) of the full universe of POTWs, our estimates of risk may be
impacted.
More serious uncertainties are implicit in assumptions required for modeling the transport
of pollutants through the environment. As with almost any attempt to estimate aggregate health
risks from exposure to pollutants in the environment, this analysis is limited by significant
uncertainty in many of the key input parameters and the] use of mathematical relationships to
perform the calculations. To reduce the analytic problem to tractable proportions, we are forced
to rely on simplifying assumptions, many of which are based on incomplete or imperfect data.
In particular, our assumptions for human behavior represent gross simplifications of the near-
infinite inter-individual variety of human behavior. For example, all exposed persons are
assume to consume the same lifetime-average mix and quantity of foods in their diet. Perhaps
more importantly, this analysis ignores mobility and time spent indoors or outdoors when
estimating potential exposure from inhalation. In other words, it is assumed that each exposed
individual resides in the same location and breathes outdoor air 24 hours per day for his or her
entire lifetime.
Significant uncertainty surrounds values selected'for characterizing soil types, uptake
rates, human behavior, and other types of parameters required for the analysis. Where possible,
the analysis gives preference to average or expected values for these parameters, but in some
cases (i.e., the groundwater, surface water, and vapor pathways of exposure from surface
disposal and land application) we rely on certain reasonable worst-case assumptions to provide
upper bound estimates of exposure and risk.
Once we have derived estimates of human exposure, we use dose-response relationships
to predict impacts on human health. As summarized above, this step is highly uncertain, and
generally involves the extrapolation of results from animal experiments (usually conducted at
relatively high doses) to humans (at significantly lower doses). For cancer, the values used to
estimate health risk are based on upper confidence limits from a conservative model of dose-
response, and are likely to over-estimate true risks.
Another limitation is that this analysis ignores certain pathways for possible human
exposure. It has focussed on those pathways of potential exposure expected to be most
1-25 :
-------
svrfte*. however, risks from the pathways that were not assessed cannot necessarily be ruled
ou For example the analysis does not consider potential risks from indirect ensure to
SSlSfrl?1^ I mCT??rS
-------
2. LEAD AND CADMIUM
* • ' '
As explained in Chapter 1, this analysis compares exposure levels to Risk Reference
Doses (RfD) for non-carcinogenic sludge contaminants. For lead and cadmium, additional
techniques are used that provide a more detailed estimate of potential health risks from human
exposure. This section discusses those techniques.
2.1 ESTIMATING HEALTH EFFECTS FROM LEAD EXPOSURE
Scientific understanding of health effects from environmental exposure to lead is more
.advanced than that for most other, contaminants considered in this analysis. Adverse health
effects from lead exposure have Ibeen recognized for many years, and pathways of human
exposure have been examined in detail by numerous studies. In addition, epidemiological
studies, both cross-sectional and longitudinal, have provided convincing evidence of association
between levels of lead exposure and the incidence of various diseases. They also provide
numerical estimates that can be used to quantify health risks from lead exposure. Animal studies
provide details of the physiological processes involved. The U.S. EPA Office of Environmental
Criteria and Assessment Office has assembled much of the available data for lead exposure and
health effects into a single document: Air Quality Criteria for Lead. Vols. I-TVr (U.S. EPA
1986c). That work provides the primary source of data and assumptions used in this analysis'
and will be referenced as the "Criteria Document" or "CD1? hereafter. '
Assumptions and Data
This report will not attempt to duplicate the extensive discussion of exposure and health
effects provided by the CD. Instead, it will briefly outline ttie major assumptions used for the
present analysis, with frequent references to that source. These assumptions can be loosely
categorized into four groups: (1) background exposure to lead from other sources and the
resulting distribution of lead "body burden" in the U.S. population, (2) absorption or uptake of
lead into the body as a result of additional exposure, (3) the manner in which additional exposure
to lead from sludge shifts the distribution of blood lead levels in exposed populations, and (4)
the relationship between blood lead concentrations and detectable health effects.
2.1.1 Background Exposure
Humans are exposed to lead through multiple environmental pathways. Anthropogenic
sources appear to dominate human exposure, so humans living in remote areas evidence
significantly lower levels of exposure than those living in urban areas of modern societies (CD,
Sections 11,13). While direct measure of an individual's exposure to lead through these
numerous environmental pathways is difficult, internal lead exposure levels can be measured
through samples of any one of several biological tissues, including blood, teeth, and bone. For
this analysis, levels of lead in human blood will be used as the measure of internal lead exposure
in human populations. Human blood lead levels (or PbB) are typically expressed in micrograms
2-1
-------
of lead per deciliter of human blood G*g/dZ). Blood lead in daily equilibrium with other
MZtT!*5 ? aduihumans aPl>ears to have a half life of 25-28 days (CD, Section 10) so
bloodjead levels are best interpreted as an indication of an individual's recent level of exposure
i
Since the physiological response to incremental lead exposure depends to a large extent
on an mdmduaTs existing body burden of lead from othS source^ assumptions atou
background distributions of blood lead levels in the United States are central to an/a^to
estimate expected I be^th risks from additional exposure. The best data currently avaSfo?
?^ I ^J^SS"" ** pK)Vided by *»• !Second National Health and Nutrition
Examination Sdidy (NHANES II), which was conducted from February, 1976 to February 1980
and included the sampBng of blood from 16,563 individuals aged 6 months to 74 ySL ftom
tiiroughout the United States. These data suggest that blood lead levels are best approximated
by log-normal distributions, with geometric means and standard deviations that differ according
to sex, race, and degree of urbanization (CD, Section 11.3.4). Based on NHANES H, the CD
reports mean Wood lead levels of about 15 ^g/dl for children aged 6 months to 5 years (of all
races) about 11 ^g/d I for women of age 18-74 years (all races), and 15.6 and 18.1 Mg/dl for
males (white and bkck, respectively) of age 18-74. MJ/UHWT
use of lie NHANES H data in estimating the current distribution of blood lead
levels requires an additional consideration: levels of lead exposure in the United States (and
STSSS^S°
years spanned by the NHANES n study, a strong association can be
detected between the use of leaded gasoline and average blood lead levels (CD, Section 1136)
and researchers have used this association to predict changes in blood lead due to the gasoline
lead phasedown. Regression coefficients derived by Joel Schwartz of the U.S. EPA Office of
fo!S^alysis have been USed by tiie U-S- ^A Office of A* Quality Planning and Standards
(OAQPS) to project average blood l
-------
Table 11-9 of the CD reports corresponding values of 1.34 to 1.39 for adult women, and 1.37
to 1.40 for adult men. The rounded ^value of 1.4 has been used in this analysis for both adults
and children.
2.1.2 Absorption and Uptake of Lead
The rates at which lead is absorbed from environmental media into human tissue depend
on the environmental medium involved, the age and sex of the exposed individual and other
factors. Predicting health effects from lead in sludge requires consideration of exposure through
air, water, and dietary pathways, and assumptions about intake from each of these sources. As
with the distribution of blood lead levels in U.S. populations, assumptions used for this analysis
are drawn primarily from the CD. These are summarized in Table 2-1 and are outlined briefly
below. ;
Air. For exposure to lead emitted by sludge incinerators, the primary pathway of
concern is the inhalation of lead from ambient air. The relationship between concentrations of
lead in ambient air and blood lead concentrations has been evaluated by a variety of
methodologies. These include experimental studies of adult volunteers, as well as
epidemiological studies of different populations of children and adults. Section 11.4.1 of the CD
reviews more than a dozen studies relating blood lead to ambient air exposures, and reaches the
following conclusion about the slope of blood lead to ambient air (0):
(1) The experimental studies at lower air lead levels, 3.2 /tg/m3
or less, and lower blood levels, (typically 30 /tg/dl or less,
have linear blood lead inhalation relationships with slopes
/S of 0-3.6 for most subjects. A typical value of
1.64+0.22 may be assumed for adults.
(2) Population cross-sectional studies at lower air lead and
blood lead levels are approximately linear with B of 0.8-2.0
for inhalation contributions.
(3) Cross-sectional studies in occupational exposures in which
air lead levels are higher (much above 10/ig/m3) and blood
levels are higher (above 40/tg/dl) show a much more
shallow linear blood lead inhalation relation. The slope B
is in the range of 0.03-0.2.
(4) Cross-sectional and experimental studies at levels of air
lead somewhat above the higher ambient exposures (9-36
/ig/m3) and blood lead levels of 30-40 /ig/dl can be
described either by a nonlinear relationship with decreasing
slope or by a linear relationship with intermediate slope,
approximately B = 0.5.
2-3
-------
Table 2-1
Estimated Intake Slopes:
Increment in Blood Lead Concentration pet Unit of Exposure
Air (/tg/dl per /tg/m3)
General Atmosphere
Adults*
Children
Point Source
Adults"
Children
Dietary and Drinking
Otg/dl per /tg/day)
Adults
Children
Low Estimate
1.8
i ie
•
2.3
1.9'
0.03s
0.15h
Middle Estimate*
1.8
•4 f A
1.5°
3.5
: 1.9f
0.04*
0.2*
High Estimate
1.8
5.0*
4.6
5.04
0.06^
0.2^
. .
"From Cohen (1988).
'Derived from Point Source estimate.
"From Criteria Document (p. 13-21). Includes soil.
Ttenved from General Atmospheres estimate.
Tram Criteria Document (p. 13-21).
Trom Criteria Document (p. 13-22, 13-23)
hFnom this report.
Midpoint.
2-4
-------
(5) The blood lead inhalation slope for children is at least as
steep as thai: for adults, with a median estimate of 1.92
from three major studies.
(6) Slopes which include both direct (inhalation) and indirect
(via soil, dust, etc.) air lead contributions are necessarily
higher than those estimates for inhaled air lead alone.
Studies using aggregate analyses (direct and indirect air
impacts) typically yield slope values in the range 3 to 5,
about double the slope due to inhaled air lead alone.
The calculations of change in blood lead level per change in air requires three steps: (1)
estimating deposition of inhaled lead, (2) estimating absorption, and (3) estimating the
incremental change in blood lead per unit of lead absorbed. As shown in Table 2-1, the present
analysis uses air intake slopes that are roughly consistent with the ranges suggested by the CD,
if indirect effects from deposition are excluded. For children, however, slopes vary with the
distance of an child's residence from the source of lead emissions. The analysis distinguishes
"generalized" atmospheres from those in areas close to sludge incinerators because of expected
significant differences hi particle size distributions.
The U.S. EPA Office of Air Quality Planning and Standards (OAQPS) has estimated
differential rates of lead absorption according to distance from a lead smelter (Cohen, 1987a).
Larger lead particles predominate at distances within 2-5 km of lead smelters, and are more
readily deposited and absorbed in the tracheo-bronchial regions than smaller particles. For the
present analysis, it has been assumed that similar distributions in lead particle size are found
within 5 km of sludge incinerators. At greater distances, large lead particles are assumed to be
less common. As described below, these different absorption estimates were applied to lead
uptake slopes to yield lead intake estimates that vary with distance.
Based on particle size distributions and lung disposition data OAQPS has found that on
average, 26-42 percent of inhaled airborne lead particles deposit in the respiratory tracts of
adults in residential locations not near a point source of lead. Differential absorption rates can
be applied to the deposition in each respiratory region to estimate that 15-32 percent of inhaled
lead is absorbed by adults in "generalized" U.S. atmospheres. For submicron particles which
dominate general atmospheres, OAQPS concluded that deposition in the lungs of a two-year-old
child is approximately 1.5 times higher than that in the lungs of adults. Using this factor to
adjust the adult absorption estimate, they conclude that 25-56 percent of inhaled lead is absorbed
by children in general atmospheres! (Cohen, 1987a).
OAQPS used the same methods to estimate total respiratory absorption of inhaled lead
particles for individuals living neat point sources of lead emissions. The expected particle size
distributing was taken from Sledge (1987) and deposition efficiencies were taken from the
Criteria Document for paniculate matter (PM) and sulfuric oxides (SOJ. By combining these
data, OAQPS estimated that near point sources of lead emissions, adults absorb approximately
38 percent of inhaled lead, and children absorb approximately 42 percent.
2-5
-------
These absorption efficiencies were applied to uptake slopes to obtain lead intake
estimates. For children, the CD reports that U.S. EPA analyses of three population studies
(Yankeltfa/., 1977; Roels etal., 1980; Angle and Mclntire, 1979) suggest toe median blood
lead mcrease is approximately 1.92 Mg/dl per /zg/m3 of inhaled air lead. All of these studies
involved children living in the vicinity of lead smelters, so they are more applicable for children
living near emission sources than for children living in generalized atmospheres. If it is assumed
that children living near sources of lead emissions absorb 42 percent of inhaled lead and children
living at various distances from these sources are assumed to absorb 25-45 percent, then it can
be concluded that the appropriate slope for children in general conditions is (25/42 x 1.92) to
( £?• X '?2) °l 11'1'2'0 "g/dl Wood lead P61 ^/m3 increment in air lead concentration. A
midpoint value of 1.5 has been used for the present analysis.
These slope estimates delitxjrately exclude indirect exposure to incinerated lead via
deposition to sod and subsequent ingestion through food, ;soil, or drinking water The CD
reports that inclusion of these indirect pathways results in estimated slopes of 3-5 ag/dl
increment in children's blood lead per ^g/m3 increment in air lead concentration For
consistency with exposure analyses for other sludge constituents, the present analysis ignores
these indirect pathways of exposure for lead.
^^818^0^8^
"™ °Pes va]lues
£? ?/? ? "S^™ f °Pes va]lues m *e range of 1.3-2.0 Mg/dl per Mg/nf with a weighted
slope of 14 (which is also the unweighted midpoint of the 0.8-2.0 range listed above)
Adjusting that value by a factor of 1.3 to account for the resorption of lead from bone tissue
%*^m% ^f ^ 1983)' °AQPS ^^rived an adjusted slope estimate of 1.8 ug/dl
blood lead per /tg/m3 increment in jiir lead concentration (Cohen, 1988).
In contrast to the available date for children, the data upon which these slopes are based
were not confined to individuals living near lead smelters. , As a result, they are likely to be
representative of general conditions in the U.S. and perhaps less appropriate for modeling lead
absorption at locations near to emission sources (where particle sizes are likely to be larger)
Using the absorption rates derived by Cohen (1987a), the slope estimate above can be adjustoi
for higher adult absorption rates near sludge incinerators. This step yields an adjusted slope of
about 3.5 Mg/dl per /*g/m3 in air lead concentration,1 a value somewhat higher than the range
of estimates for general atmospheres quoted from the CD.
For comparison, deposition and absorption rates derived by Cohen (1987a) can be combined
with uptake rates relating changes in blood lead levels to increases in the amount of absorbed
lead Based primarily on results from tracer studies by Rabinowitz (1976, 1977) and Marcus
(1988) estimates that each jig/day of lead absorbed by adults results in an approximate 0 4 ug/dl
increment in blood lead. If this estimate is combined with the finding of Cohen(1987) that adults
absorb approximately 38 percent of inhaled lead when residing near lead emissions, and with
the assumption that a typical adult inhales approximately 20 m3/day (U.S. FJA, 1986c) it can
be concluded that each pg/m3 of lead in ambient air would result in an increment of 20x0.38x0 4
or about 3 /tg/dl in blood lead for these individuals, a conclusion in approximate agreement with
the above estimate.
2-6
-------
2.1.3 Dietary or Drinking Water Pathways
«
The CD reports that typical absorption rates for ingested lead are 10 percent for adults
and 25-50 percent for children. Combining these estimates with uptake rates of 0.4 for adults
and 0.5 for children (from Marcus, 1988), yields an intake slope of 0.04 for adults and 0.13-
0.27 (midpoint 0.2) for children. By comparison, the CD reports intake slopes of 0.04 for
adults and 0.16 for infants. Estimation of intake from drinking water is based on the same
results. It is assumed that the aveitage adult consumes. 2 liters of drinking water per day and the
average child consumes 1 liter daily. Lead consumed through drinking water is assumed to have
the same rates of absorption and uptake as lead consumed'through the diet.
2.1.4 Shifting the Blood Lead Distributions
The preceding section of this report discusses the rates at which exposure to additional
lead can be expected to affect the concentration of lead in an individual's blood. More directly
pertinent to the estimation of health risks from sludge is the question of how additional exposure
to lead might affect the distribution of blood lead levels in an entire exposed population. Since
it has been assumed that the distribution of blood lead values in a given population is
approximately log-normal, and since a log-normal distribution is completely characterized by its
geometric mean and geometric standard deviation (GSD) this question is equivalent to asking
how additional lead exposure affects the geometric mean arid GSD of blood lead in the exposed
population.
A log-normal distribution is described by x=efll+ai) where e is the base of the natural
logarithm function, p is the natural log of the distribution's geometric mean,
-------
Table 2-2
Sample Calculation:
Lead in Drinking Water
Geometric Mean Blood Lead Without
Exposure from Sludge (jtg/dl)
Increment to Water Concentration (mg/J)
Incremental Lead Ingested Oig/day)
Average Increment to Blood Lead Level
Geometric Mean Blood Lead After
Exposure from Sludge (pg/dQ
Adults
4.59
0.0002
0.4*
0.0i6«
4.606
Children
5 41
0.0002
0.2b
0.04d
5.45
'Assumes adults ingest 2 liters of water daily.
"•Assumes children ingest 1 liter of water daily.
'Assumes intake slope for adults of 0.4 Mg/dl PbB per Mg/day of lead ingested
Assumes intake slope for children of 0.2 /*g/dl PbB per /tg/day of lead ingested
The geometric standard deviation for the distribution of blood lead is assumed to remain
constant at 1.4 jtg/dl. *^maui
2-8
-------
2.1.5 Health Effects
*
This section discusses the health effects quantified in this analysis. The model described
was originally developed in-support of the Office of Water's proposed regulation for sewage
sludge. Later, it was refined and used in support of drinking water regulations by U.S. EPA's
Office of Ground Water and Drinking Water (OGWDW), as described in U.S. EPA (1990d).
Categories of potential health benefits are summarized iin Table 2-3; those health effects
quantified in this analysis are highlighted.
Background. U.S. EPA bus conducted numerous studies on the health effects associated
with lead exposure. In a pioneering study (U.S. EPA. 1985c), Schwartz et al. quantified a
number of health benefits that would result from reductions in the lead content of gasoline. The
work was extended by U.S. EPA's analysis of lead in drinking water (U.S. EPA, 1986d) and
by a U.S. EPA-funded study of alternative lead National Ambient Air Quality Standards (U.S.
EPA, 1987c). Despite this substantial research, much uncertainty remains. Many categories
of health effects from lead exposure cannot be quantified - Credible dose-response functions are
not yet available. There is also uncertainty regarding the significance of many of these health
effects. It is not clear whether the estimates provided in this chapter overestimate or
underestimate the actual values.
Threshold Approach. Based on information from the CD, thresholds were selected
above which adverse health effects from lead were considered possible: 7 /*g/dl for men, and
10 /ig/dl for women and children. Using the estimated background distributions of blood lead
levels discussed above, the number of persons exceeding these thresholds was estimated. Next,
using estimates of potential exposure to lead from sludge, these distributions were shifted. From
these revised distributions, the numbers of men, women, and children exceeding the selected
thresholds were calculated again. Ely subtracting the number of persons exceeding the thresholds
before and after exposure from sludge, the incremental number of persons at risk of health
effects from lead in sludge could be calculated.
Children with IQs Less tlbian 70. Based on encoded expert opinion, Wallsten and
Whitfield (1986) provide estimates of expected percentages of children with IQs below 70 for
each level of population mean blood lead. For this analysis, the opinions of the experts were
averaged.
Health Benefits to Adult Men. With the availability of the Second National Health and
Nutrition Survey (NHANES IT) several studies, including Pirkle et al., (1985) and Harlan et al.
(1985), have found a statistically significant relationship between blood lead and hypertension
in adult males. In particular, Piikle et al. found that blbod lead levels were a significant
predictor of blood pressure in adult white males. They found that each increase of one log unit
in blood lead in males of age 40-59 could be associated with 3.954 points of diastolic blood
pressure and 8.436 points of systolic blood pressure. These: relationships held when blood lead
was evaluated in a regression with all factors previously known to be correlated with blood
pressure, and 87 additional variables representing combinations of every dietary and serologic
variable in the NHANES H survey. Results from a large-scale study of British men
(Pocock et al., 1985) are consistent with these results.
2-9
-------
Table 2-3
*
Potential Health Benefits from Reducing Exposure to Lead
Men
Hypertension in adult men
Myocardial infarction, stroke and death in men of ages 40-59
Myocaidial infarction, stroke, arid death hi men of other ages
Cancer
Women
Hypertension, myocardial infarction, stroke, and death
Fetal effects from maternal exposure, including diminished IQ, decreased gestational age,
and reduced birth weight
Possible increases in infant mortality resulting from maternal exposure
Cancer
Children
Interference with growth
Reduced intelligence
Impaired hearing, behavioral changes
Interference with development of Peripheral Nervous System
Metabolic effects, unpaired heme synthesis, anemia
Cancer
These health effects have been quantified for this analysis.
2-10
-------
Reduced Incidence of Hyj tertension. One of the authors of the Pirkle et al. study, Joel
Schwartz, has continued to investigate the relationship between blood lead and hypertension in
support of the benefit analysis for the phasedown of the lead in gasoline (U.S. EPA, 19856).
Using multiple logistic regression, Schwartz derived a function that could be used to predict the
incidence of hypertension in white males of ages 40-59 from! individual blood lead concentrations
and other variables. Results of the regression analyses are reported in U.S. EPA (1985c).
Schwartz used this equation to simulate changes in the probability of hypertension for each
observation in the NHANES data as a result of changes in each observation's blood lead level.
By aggregating results, potential impacts of the gasoline lead phasedown were estimated.
Similar techniques were later used in U.S. EPA (1986d) for estimating benefits from reducing
lead in drinking water.
U.S. EPA (1987c) used results from Schwartz's multiple logistic regression model to
derive a univariate logistic function which predicts the probability of hypertension (diastolic
pressure above 90 mm Hg) based on blood lead levels alone:
= [1 + exp( -(-2.744 + 0.793 (In P&B,))) J'1 -
[1 -K exjK -(-2.744 + 0.793 (In PbBJj) }~l
where:
APrtTHYP) = the change in the probability of hypertension,
blood lead level before some change, and,
blood lead level after some change.
U.S. EPA's Office of Air Quality Planning and Standards ;derived this univariate function by
replacing each of the independent variables with corresponding mean values from the
NHANES n sample. Since the logistic equation is non-linearj this method of reducing the
Schwartz equation to a single independent variable is likely to introduce error into the estimation
process, but is thought to provide a reasonable approximation of the effect of blood lead on
blood pressure. The function derived by OAQPS (which is graphed in Figure 2-1) provides the
basis for estimating cases of hypertension associated with lead in sludge.
As reported in U.S. EPA (1985c) and U.S. EPA (1986d) these regression equation results
are based on a sample population of white males of ages 40-59. More recently, Schwartz (1988)
nas tested the relationship between blood lead and blood pressure with a broader population,
including all males (black and white) from 20 to 74 years of age. He found that the relationship
was consistent over the entire population tested. For the present analysis, it is assumed that the
univariate equation derived by U.Si. EPA (1987c) can be applied to all adult males (black or
white) between 20 and 74 years of age. By combining the! dose-response curve with pre- and
post-regulatory blood lead distributions, we calculate the number of cases of hypertension that
are associated with lead in sewage sludge disposal practices.
Reduced Incidence of Coronary Heart Disease Events. Serious health consequences
often result from hypertension. Several large epidemiological studies have shown that elevated
2-11
-------
0.45
0.4
0.35
0.3
ff 0-25
•8
£•- 0.2
0.15
0.1
0.05
0
B!gun;2-l
Probability of Hypertension vs. Blood Lead
4 68 10 12 14 16 18 20
Blood Lead (ug/cfl)
2-12
-------
blood pressure increases the risk of coronary heart disease (CHD). The Pooling Project (1978)
and the Framingham studies (Shurtleff, 1974 and McGee and Gordon, 1976) are well known
studies that estimate the incidence of*coronary disease as a function of blood pressure, smoking,
and other risk factors. Levy et al, (1984) have shown that the risk coefficients from the studies
have accurately predicted the 'decline in coronary incidence iin the 1970s associated with reported
declines in blood pressure, smoking, and cholesterol levels over the same period.
Unfortunately, none of these prospective studies recorded measures of blood lead levels
for the men investigated. It is therefore impossible to use their data directly to derive estimates
of the relationship between blood lead levels and heart attacks, strokes, or premature death.
However, indirect methods can be used to estimate dose response relationships. For each health
endpoint, two steps are involved. First, estimates of blood lead levels in a population to predict
its expected mean diastolic blood pressure are used. Next, this blood pressure estimate is used
in conjunction with results from the Framingham and Pooling Project studies, to derive estimates
of the incidence of heart attacks, strokes, and death. To calculate changes hi these health
._ endpoints, all calculations are repeated for a population wjhase.jmemjblood lead. Level has
changed. This value is subtracted from the baseline to estimate health consequences of
increments in lead exposure.
Estimating Shifts in Population Mean Blood Pressure as a function of Blood Lead
Level. As mentioned above, PirMe et al. (1985) estimated that each log unit increase in an
individual's blood lead level could be associated with an expected increase of 3.954 millimeters
of diastolic blood pressure. Schwartz performed additional regressions on the same data to
derive a coefficient of 4.609 (U.S. EPA, 1985c). As reported in U.S. EPA (1986d), Schwartz'
results were later challenged for his having failed to control for the 64 sites involved hi the
survey. Subsequent work by Schwartz (1986b) repeated the regressions with the addition of
dummy variables for each of the isampling sites hi the survey, to yield a revised coefficient of
2.74 for diastolic pressure, an approximate 40 percent decrease from that reported in U.S. EPA
(1985c). More recently, Schwartz (1988) reports having used a random effects model to test
the effect of site variation on the regression results, and finds that regression coefficients for the
log of blood lead are reduced by about 25 percent from original estimates for systolic and
diastolic pressure, to yield a somewhat higher coefficient than the 2.74 value reported in U.S.
EPA(1986c).
U.S. EPA (1987c) recommends using the 4.609 coefficient from U.S. EPA (1985c) to
construct the equation:
ABP = 4.609 (hi PbB^ - PbB2) (2-1)
where:
ABP = the change in blood pressure expected to result from a change in blood
lead as a result of regulatory controls,
PbB, = blood lead level before regulatory controls, and
PbBj = blood lead level after regulatory controls.
2-13
-------
pressure is linear with respect to all other explanatory
aT^LTZ ISST"""" mu"*e modelof as-BA (1985c) is S*- 2
site-adJusted coefficient of 2.14 is used (from U.S. EPA
fi ;raSSUie With respect to cfaanSes m blood lead levels. Results
ABP ••= 2.74 (In PbB^ - In PbBJ (2-2)
Using Shifts in Blood Pressure to Estimate Changes in Rates of CHD Events Once
sinan exi»^iv™iafi™'c meah bloodpressure have been predicted, results from the"
. . . T _.., studies can be used to predict changes' in other health
endpoints. In Piride etal (1985), U.S. EPA (1985c), and U.S. EPA (1986d), regression
described above were used to shift the expected blood pressure for each fafi^ShSSi
^J£^^ZZ^*^-™««- ^ ^^
survey. sng tese shifted estimates to
N^ ^Ktr11 Obsf™*™:*.^ to ^king, age, and serum cholesterol frT
NHANES H^e authors applied Iqgistic models from the Pooling Project and Framineham
studies to predict each sample individual's odds of a heart attack steofc- or d^ T
predicted **•-* expected - °f ^fi^fiS. 2
U.S. EPA (1987c) offers a similar, but simplified technique for perfo
calculations. IHey begin by applying Equation 2-1 to predict a shift m L^LS
SsTofm^dT8? ^"^.Next, they used a simplified version of the mul
6 aU)Ilfied
they used a simplified version of the multiple logistic
univariate equation describing the probability of a CHD
*""""""" As with the logistic eauation for
)ased on these studies to derive a substitute equation bf a single explanatory variable:
blood pressure. For first coronary heart disease events, the resulting equrtion is:
= [1 + exp(-(-4.996 + 0.030365 BP,))]'1 -
(2-3)
[1 + exp(-(-4.996 + 0.030365 BP,)))]'1
where:
APr(CHD) = the change in the probability of occurrence of a CHD event,
B?! - mean diastolic blood pressure before regulatory controls, and
BP2 - mean diastolic blood pressure after controls.
2-14
-------
By combining Equation 2-3 with Equation 2-1, OAQPS estimates changes in the rates of first
cardiac events (over a 10 year follow jup period) based on changes in population mean blood lead
levels.
«
To what extent does this simplification introduce error into the estimation process? Since
the logistic equation used to predict first CHD events is non-linear with respect to age, serum
cholesterol and smoking, the substitution of mean values for these variables will not necessarily
result in an equation equivalent to the one developed from the Pooling Project data. To evaluate
the possible error introduced by this procedure, we compared dose response curves generated
by Pirkle et al. (1985) to results generated with the OAQPS methodology (using the 3.954
coefficient from Piride et al. for blood pressure versus the log of blood lead). From this
comparison, it seems that results from the two methods are quite similar, suggesting that the
practical advantages of the OAQPS methodology outweigh any losses to accuracy introduced by
this analytic step.
Therefore, the OAQPS methodology was adopted with one modification: to estimate
shifts in diastolic pressure, Equation 2-2 is substituted for Equation 2-1. Annual cases are
predicted by dividing the results by 10 years to derive a dose-response curve. Using this
function, the number of additional CHD events expected to result from exposure to lead in
sludge is estimated.
One limitation to the use of Pooling Project data to predict rates of CHD events is that
the regression models from those studies were restricted to white middle-aged men (about a third
of the entire adult male population). If logistic regression coefficients were available for a
broader population of adult males, then blood pressure - blood lead regression coefficients
derived by Schwartz (1988) for males 20-74 could be used to estimate changes in rates of CHD
events for this larger population. Recent research suggests that the same relationship holds for
black men as well.2 For this analysis, changes in rates of CHD events are predicted for all
males of ages 40-59. The estimates ignore reductions in rates of CHD events that might result
from men of other age groups.
Reduced Incidence of Stroke. U.S. EPA (1987p) used similar methods to derive
univariate equations relating the incidence of stroke and death from all. causes to changes in
population mean blood lead levels (through changes in diastolic blood pressure). For strokes,
they use results from Shurtleff (1974) as listed hi Tables 2-4 and 2-5. These univariate logistic
equations are applied to estimates of shifts in population mean blood pressure (derived with
Equation 2-2) to yield changes in rates of stroke as a result of changes in lead exposure. For
estimates of effects of reduced lead exposure on rates of stroke, the OAQPS methodology is
modified only slightly, by using a site-adjusted coefficient (Equation 2-3) to estimate changes
in blood pressure. Combining the regression equations of Shurtleff (1974) with Equation 2-3
yields the dose-response curve used in the current analysis: This function is entered into the
exposure model to estimate changeis in the incidence of stroke due to lead in sewage sludge.
2See Schwartz (1988).
2-15
-------
Table 2-4
Logistic Regression Relating Blood Pressure
.^ to the Probability of Initial
Cerebrovascular Accident in White Men Aged 45-74
Variable Coefficient t-Statistic
Constant - -8.58889 'I
Diastolic Blood Pressure 0.04066 5 77
* Not reported
Source: Shurtleff (1974), as taken from U.S. EPA (1987c). p. 6-13.
Table 2-5 •
Logistic Regression Relating Blood Pressure
to the Probability of Initial
Atherothrombotic Brain Infarction in White Men Aged 45-74
Coefficient t-Statistic
Constant -9.95160
Diastolic Blood Pressure _ 0.04840 5 16
* Not reported
Source: Shurtleff (1974), as taken from U.S. EPA (1987c). p. 6-13.
2-16
-------
Reduced Incidence of Premature Death. Data from the Framingham Study have also
been used to estimate the relationship between blood pressure and rates of death from all causes.
Pirkle etal. (1985), U.S. EPA (1985C), and U.S. EPA (1986b) use logistic equation coefficients
from the Framingham Study to predict changes in the rates of death as a function of changes in
blood lead levels. '»
U.S. EPA (1987c) used population mean values for serum cholesterol and smoking to
reduce results from the Framingham Study to an equation in one explanatory variable:
APr(Jtf.R7) = [1 + exp(-( -5.3158 +
(2-4)
[1 + exp(-(-5.3158
where:
APr(MRT) = the change in probability of death,
B?! = the level of diastolic blood pressure before some change, and
BP2 = the level of diastolic blood pressure after some change.
As with estimates for CHD, dose^response curves -generated by this simplified equation. are
compared with those using the multi-variable model and individual data observations as reported
in Pirkle et al. Once again, the results compare favorably with curves reported from the original
study (if a coefficient of 3.954 is substituted into Equation 2-1). For estimates of death caused
by lead in sludge, therefore Equation 2-2 is combined with Equation 2-4 to generate the dose-
response curve shown in Figure 2-2.
Health Benefits for Women. None of the methods outlined thus far includes
consideration of possible health consequences of women's exposure to lead in sludge.
Nevertheless, at least some available evidence suggests the possibility of such benefits. Recent
expanded analysis of NHANES n by Schwartz (1990) indicates a significant association between
blood pressure and blood lead in women. Another study, by Rabinowitz et al. (1987), has found
a small but demonstrable association between maternal blood lead and pregnancy hypertension
and blood pressure at time of delivery. Finally, a recent study of NHANES n data by
Silbergeld et al. (1988) suggests that accumulated lead stores in the bone tissues of women may
be mobilized into blood during conditions of bone demineralization associated with pregnancy,
lactation and osteoporosis. The authors note that "lead may interact with other factors in the
course of postmenopausal osteoporosis, to aggravate the course of the disease, since lead is
known to inhibit activation of vitaimin D, uptake of dietary calcium, and several regulatory
aspects of bone cell function." No quantitative relationship has yet been established, however,
between lead stores in women and postmenopausal health ehdpoints. For kck of sufficient data
to quantify these and other potential impacts of lead exposure on women's health, this analysis
does not attempt to quantity health benefits from reductions in women's exposure to lead.
2-17
-------
2.2 ESTIMATING HEALTH EFFECTS FROM CADMIUM
,,ooc ,MeJhods for estimating health risks from cadmium .exposure are outlined in U S EPA
(1985a). In that document, human health effects from cadmium exposure are described as
IOJU.O Yr S •
Damage to the kidney's ability to reabsorb blood protein is the known
(non-carcinogenic) effect having the lowest exposure threshold. Increasing
degrees of cadmium induced renal tubular dysfunction are manifested in
Bj-microglobulin proteinuria, general proteinuria, aminoaciduria, and
glycosuria, in order of increasing severity. Effects on bone and mineral
metabolism have accompanied kidney damage in sever (sic) cases as found
in the Itai-Itai or "Ouch-Ouch" disease in Japjm (U.S. EPA 1979, 1980).
Elevated B2-microglobulin excretion is not equivalent to clinically
significant proteinuria. Without continued high exposure to cadmium,
there is little evidence of either a progression of severity of kidney
dysfunction, or a significant shortening of life expectancy. Nevertheless,
while some elevation of Vnucroglobuliri excretion appears to be a
relatively benign condition, it is usually taken as the threshold health
effect in setting ambient cadmium criteria (U;S. EPA 1979, 1980).
As with lead, data are available describing potential health effects as a function of the
burden of cadmium accumulated in an individual's body. Unlike lead, the depuration rate of
cadmium 1S relatively slow, with a half life of 18-38 years, and the commonly used measure of
cadmium accumulation is the concentration in an individuals's kidney cortex. U.S. EPA (1985a)
reports that 25 percent of inhaled cadmium enters the blood stream, as compared to 5-6 percent
of ingested cadmium. The authors continue:
B2-Microglobulin proteinuria may occur when the concentration in kidney
cortex reaches approximately 200-400 pg/g (wet weight), although
individual susceptibilities may fall outside this range. At the present time
a kidney concentration of 200 pg/g is the most widely accepted estimated
of the critical threshold (Ryan et al., 1982).
This concentration is estimated to result from a daily retention (absorption)
rate of 10-15 /tg/day over a 25-50 year period (U.S. EPA 1979, 1980).
If 12 jtg/day is taken as the absorbed dose that will produce a Mdney
cadmium level of 200 pgfday over a 25-50 year period, then at a 6
percent absorption efficiency this corresponds to a gross ingestion of 200
Mg/day, a value sometimes cited as a threshold (Commission of the
European Communities, 1978).
As with lead, the present analysis combines information describing background levels of
tissue concentration, estimated exposure from sludge, and estimated intake slopes to calculate
the extent to which the background distribution of a population's burden of cadmium will shift
2-18
-------
as a result of additional exposure to cadmium from sludge. Given this estimated shift, the
number of persons crossing a spxafied kidney cadmium "threshold" (200/tg/g) is used to
represent the number of persons at risk from sludge disposal.
•
Based on the information quoted above, 200 /*g/day of cadmium ingested over 25-50
years results in an eventual kidney cortex cadmium level of about 200 /*g/g. If the relationship
between kidney cortex concentration and ingested cadmium is approximately linear, then it can
be reasoned that each pg/day of lifetime average cadmium ingestion regulates in about 1 jtg/g
in eventual tissue concentration. Since only 5-6 percent of the ingested cadmium is absorbed,
this means that each /tg/day of absorbed cadmium contributes about 16-20 pg/g to the eventual
burden. These estimates were used to translate incremental exposure to cadmium from sludge
into estimated average increments in. kidney cadmium concentrations for the exposed populations.
An estimate of the distribution of the background concentration of cadmium in tissues
was obtained from U.S. EPA (1985a). This report describes autopsy studies of 93 men, all
more than 30 years in age, that revealed an approximately log-normal distribution of kidney
cadmium concentrations, with the following characteristics:
Geometric Geometric
Mean(fig/g) Standard Deviation 0*g/g)
Category Number r95% Confidence Range] F95% Confidence Range]
Nonsmokers 21 15.0 [11.7-19.3] ; 1.74 [1.53-2.19]
Smokers 72 27.9 [25.0-31,2:| 1.60 [1.50-1.75]
Combined 93 24.2 [21.6-27.1] 1.73 [1.61-1.89]
The men had no known occupational exposure to cadmium.
If it is assumed that exposure to the additional quantities of cadmium from sludge does
not appreciably alter the geometric standard deviation of kidney cadmium concentrations in the
exposed population, then the number of persons exceeding the 200 jtg/g threshold of potential
health effects can be calculated before and after exposure to cadmium from sludge, and the
results subtracted to yield possible health risks from sludge!disposal.
For this report, potential health risks were estimated separately for smokers and
nonsmokers. For exposure from distribution and marketing of sludge, for which children were
modeled separately from adults, accumulated cadmium from the time-weighted childhood
exposure was added to that from adult exposure to approximate the eventual total cadmium
burden of an adult with 25-50 years of exposure. Absorption efficiency for children (through
ingestion) was assumed to be the same as for adults. Since data were not available to describe
the background distribution of kidney cadmium concentrations in adult women, it was assumed
that their distribution can be approximated by the study of adult men; the results reported in the
U.S. EPA (1985a) were used to estimate potential health risks for both men and women
combined. To the extent that these data under- or overestimate kidney cadmium concentrations
for women or younger men, these results may under- or over-estimate potential health risks.
2-19
-------
of the number of persons at risk of adverse health effects.
2-20
-------
3. INCINERATION
3.0 INTRODUCTION .
t
This chapter describes the methodology and data used to estimate human health risks
from the incineration of sludge. It reports our estimates of risks under current ("baseline")
conditions, and our estimates of the risk reductions likely to be achieved by the regulation. Only
direct exposure through inhalation of pollutants is considered; risks from indirect exposure
through the deposition of pollutsints onto soil, crops, or surface water bodies have not been
examined. Also excluded from the analysis are potential risks from the disposal of incinerator
ash. Finally, this analysis considers human health risks only, and does not account for potential
adverse effects on plants and animal life.
Section 3.1 will discuss the mathematical models and other methods used to calculate
human health risks from the incineration of municipal sewage sludge. Section 3.2 will discuss
the sources of data used for this analysis, and will provide values for the input parameters used
for the modeling. Section 3.3 will provide estimates of health risks before implementation of
regulatory controls, and estimates of the health benefits to be achieved by regulating the
incineration of sewage sludge.
3.1 METHODOLOGY
This analysis uses four steps to estimate baseline risks from incineration of sludge:
(1) estimate the rate at which pollutants are emitted from incinerator facilities;
(2) estimate the transport arid dispersion of pollutants in ambient air near incinerators,
and determine the extent to which pollutant plumes overlap;
(3) map expected, ground-level concentrations of pollutants onto human populations;
and
(4) determine the extent of human exposure to emitted pollutants and the resulting
health risks.
Based on assumptions about reductions in emissions from regulatory controls, these four steps
are repeated to estimate human health risks after the regulation is in place. The difference
between the two estimates describes the health benefits to be achieved by the regulation.
As discussed in Chapter 1, the basic strategy of this aggregate risk analysis is to assume
that the stratified sample of POTWs in the analytic component of the National Sewage Sludge
Survey (NSSS) can be used to represent the full inventory of POTWs in the U.S. In general,
aggregate risks estimated for this sample of plants are scaled by sample weights to derive
estimates of risk at the national level. For incineration, we modify this approach slightly''to
3-1
-------
exploit additional data available from other sources, and to account for the fact that some
populations will be exposed to eiriissions from more than one incinerator.
» .
3.1.1 Estimating Emissions of Pollutants
The first ftp in estimating human exposure and risk is to determine the rate at which
*?* *"? ±G StaCkS °f *«*«** **»»• TTie method for calculating
r^ ^ °f f0^' For metals' * *s based on the mass of pollutant entering
devis re effiaency of the furnace, and any operating poUution control
' ^ - 3.17x10-" C. Mp (1-iy
where:
EJP = emission rate for contaminant j at incinerator facility D (g/sec)
J = index for pollutants,
P = index for incinerator facilities, '
3.17x10^ = constaint to convert units from ;(g/yr) to (g/sec),
CJ = concentration of metal j in sludge (grams per dry metric ton, or
g/DMT),
H, = mass of sludge incinerated at facility p each year (DMT/yr), and
«JP = combined removal efficiency for pollutanty of furnace and '
control devices for incinerator/? expressed as fraction of original
contauainant retained by the furnace or poUution control devices
(dimensionless).
To calculate the average rate that metals are emitted from each facility, we first calculate
the rate at which a contaminant enters the facility, based on the feed rate for sludge (M.) and
the concentration of contaminant in the feed (q). For a given mass of a metal contaminant
entering the incinerator, some fraction will remain in the bottom ash of the furnace Of the
remainder, some is trapped by poUution control devices and the rest is emitted from the stack
To estimate the fraction of contarainant released to the atmosphere, the mass entering the
incinerator (per unit time) is adjusted for the removal efficiency of the furnace and controls OU
The resulting estimates for emissions from individual facilities represent stack emissions in unite
ot grams per second for each contaminant
For organic poUutants, predicting rates of emission is more complicated, because organic
pollutants from sludge can be destroyed within the incinerator, and other organic compounds can
oe termed in the incinerator as products of incomplete combustion (PICs). Emission rates for
organic poUutants are therefore estimated directly from results of monitoring studies for sludge
incinerators. Emissions are assumed to be determined by the type of furnace used, the quantity
of sludge incinerated, and the use of poUution control devices-
3-2
-------
EA> * 3.17xiO-8 0^ Mp (1-jy
where: * !
% = emission rate for organic pollutant j at facility/? (g/sec),
3. IVxlQ-8 = constant to ^convert units from (yr1) to (sec'1),
Ojp = emission rate for organic pollutant,/ for a unit sludge feed rate at a facility
of the same type as facility p (g/sec per DMT/sec),
Mp = mass of shuige incinerated per year at facility p (DMT/yr), and
% = removal efficiency of control for pollutant j for facility of same type as
facility p, expressed as the fraction: of pollutant retained by the control
device (dimensionless).
The mass feed rate of sludge (Mp) (converted to units of DMT per second) is multiplied
by the unit emission rate for the appropriate furnace type (O^,) (estimated with methods to be
described in Section 3.2) to obtain estimated emissions in grams per second for each pollutant
and each facility. Estimated emissions are reduced proportionately to adjust for the effectiveness
of pollution control devices operating at each facility (R^;. We assume organic emissions are
reduced only by afterburners: other pollution control devices have negligible impact on organic
emissions. Note that, unlike the calculation for metals, concentrations of organic contaminants
in the sludge are not considered when estimating expected emissions of organic pollutants.
3.1.2 Modeling the Dispersion of Pollutants in Air
Dispersion of pollutants in air is simulated with the Industrial Source Complex Long
Term (ISCLT) model (Bowers et al., 1980; U.S. EPA, 1986J) as implemented in the current
version of the Graphical Exposure Modeling System for personal computers, or PC-GEMS (U.S.
EPA, 1989d). The model describes the dispersion of pollutants as steady-state, Gaussian
plumes, aiid allows the user several modeling options. Aerodynamic downwash can be
considered if nearby building dimensions are known. Plume rise can be predicted as a function
of distance. Chemical degradation during transport can also be considered in the calculations.
Emissions can be classified as originating from either point, area or volume sources.
The ISCLT model requires several parameters describing the gas plume as it exits the
incHerator stack, as well as further information to determine how the plume is affected by
surrounding buildings and terrain. The height and inner diameter of the stack are required
inputs, and the velocity and temperature of exit gases must also be specified. The height and
effective width (square root of the area) of the building nearest the stack are needed to evaluate
the effects of potential downwash on the plume.
We model all incinerator stacks as point sources. Depending on the velocity and
temperature of exit gases, plume rise is modeled as either momentum- or buoyancy-induced; the
appropriate option is selected automatically by the program; We invoke both the downwash and
plume-rise-by-distance options, but (for lack of sufficient data) ignore the effects of surrounding
terrain. We conservatively assume that pollutants do not degrade significantly during air
3-3 i
-------
transport before reaching human receptors. This last assumption simplifies our calculations
because transport modeling for different pollutants is identical if decay is not oooSwTS
computational efficiency, the dispersion of pollutants near each facility is modeled only once
^0' CmiSSi0nS (Le" ' g/MC°f P°U^e-^erk/^of slud '
K • COBVWtod to *™nd-tevel concentrations at indivi
locations, scaled by expected emissions of each pollutant from each individual facility.
3.1.3 Mapping Dispersion and Pollutant Concentrations Onto a Unified Grid
Results from the ISCLT model are reported as dispersion ratios in units of ug/nf of
ST °Q -m T ^ Jdr P" g/S6C Of emissions- S^^te coefficients are pvded
for selected locations in the area surrounding an individual incinerator facility. The rTodel
s. We
del
US00* ™>*"»i- " ~~ "*"i'i'i"6 WJt ""i«=iaiuu lauus for
•
specify coordinates in such a way that results from the modeling
to ktegrated ^ a U0ified ma^S of **«*» ntftos for ftj
flnm nf ^ ^^^genic, threshold-acting pollutants, the potential overlapping of plumes
from multiple incinerators can affect the characterization of health riskTto tf» eWed
population. In other words, although pollution from a single incinerator might not be sufficient
£S£ *S HXP°- mdiVMUa! '? 6XCmI a ^ reference dose> additional Portion from S2
(nearby) sludge incinerator might raise total exposure to a level exceedingtfae RfD Moreover
h^th effect rektionsfaips established for lead and cadmium are non-lineal wM rrespe* ^%£
For these poUutants, summing estimated health effects from individual incineSs in Ae
national inventory wou d not necessarily provide an accurate estimate of health risks from aU
incinerators operating simultaneously.
tiKfrn, T° ,aa?U!! fOT f0^ ?"**&* of P°Uu^t plumes, we construct a unified grid
system onto which results from the modeling of individual incinerators can be integrated One
desirable characteristic for such a grid is that the cells of overlapping grids of neighboring
facilities be perfectly aligned. A simple and intuitive coordinate s^m for that gnTwoSl
to use fractions of degrees latitude and longitude to describe nodes in the grid system. The
latitude and tongitude of most incinerating POTWs is recorded in data from the 1986 NEEDS
baS^f^ ^f h d^igned ? aCCSSS meteorolo'gical and population data automatically
based on latitude and longitude coordinates. However, a complication with relying exclusively
on this coordinate system is that the distance represented by a single degree of longitude
decreases with increasing latitude; if the same dimensions in units of latitude and longitude are
used to define the area modeled near each incinerator, the widths of areas modeled for facilities
in the northern U.S. will be significantly smaller than the widths of corresponding areas for
southern facilities. If too small an area is modeled for a, particular facility, risks may be under-
counted. Conversely, if the grid spacing is too large, the model might fail to resolve rapid
changes in pollutant concentrations near incinerators.
To achieve consistent dimensions while aligning grid cells into an integrated system we
ct a transformed, two-dimensional grid system to approximate the locations of
construct __j
3-4
-------
incinerators and human populations. Any transformation from mapping on a sphere to mapping
on a flat surface produces distortion. Our goal has been to derive a simple mapping scheme that
holds this distortion to levels of little significance for model calculations. First, the latitude and
longitude coordinates of each facility are converted to spherical coordinates in a system with its
origin at the center of the earth (assumed to be a perfect sphere of radius 6452 km). Second,
the spherical coordinates are transformed to a three-dimensional cartesian grid, maintaining exact
locations. Finally, positions of each incinerator are approximated with a two-dimensional (2-D)
grid. The origin of the new grid is chosen to be the approximate center of the continental U.S.
(latitude: 37, longitude: 96). Because the distortion inherent in this step is greatest at large
distances from the origin, setting the origin at the grid center helps reduce maximum error.
To determine coordinates in the 2-D system., exact distances are first calculated along
arcs on the earth's surface between each facility and the origin of the 2-D system. Positions on
the 2-D grid are selected so these distances remain exact, A facility's ^-coordinate in the new
system, which represents distance in the North-South direction, is defined by the facility's angle
of latitude with respect to the 2-D origin. At all locations, one degree of latitude is equivalent
to approximately 113 km. This y-coordinate is used together with the true distance between the
facility and the origin to determine the Jt-coordinate, which represents distance in the East-West
direction. Distances between facilities as calculated based on the resulting 2-D grid consistently
fall within a few percent of true distances along the surface of the earth, with agreement
improving as distances decrease.
Within this coordinate system, we use two sizes of grid cell to cover sufficient area while
providing sufficient detail at small distances from each plant. As shown in Figure 3-1, we
define large grid cells with dimensions of 4 km by 4 km/ ,For simplicity, modeled locations for
plants are adjusted by a maximum of 2 km to the center of a grid cell. The grid system extends
for ten cells in each direction from the facility, for a total of 84 km in both width and length (42
km from the center to each edge of the grid, and approximately 59 km from the incinerator to
each corner). Because greater refinement is desired closer to the incinerator (where
concentrations are expected to be greatest), each large grid cell containing a facility is further
divided into 400 small grid cells (each of dimensions 200'x 200 m) with the incinerator located
at the central node (as shown in Figure 3-2). After slightly adjusting the latitude and longitude
coordinates of each actual facility and specifying receptor locations according to the grid system,
we use ISCLT to calculate dispeirsion ratios appropriate for the center of each cell in both the
large and small grid systems. The only location for which a dispersion ratio is not calculated
is the center of the large grid system, which represents the location of the incinerator. Within
the small grid system, receptor locations are specified at distances from 100-1900 m at intervals
of 200. Within the larger system, they are specified at distances from 4-40 km at intervals of
4 km. Two separate executions of-ISCLT are necessary; to calculate dispersion ratios for the
receptor locations in the small and large grid systems for each facility.
The next step is to calculate expected concentrations of each pollutant within each cell
of the grid system. To do so, we combine estimates of emissions from each facility with results
from ISCLT to calculate each incinerator's contribution to pollution within surrounding cells.
Where a cell is included in the grid system for more than one incinerator, results for each
pollutant are summed to calculate total expected concentrations. The total concentration
3-5
-------
figure 3-1
Large Grid For Modeling Incinerators
4km
3-6
-------
Figure 3-2
Facility Location in Small Grid
200m
— Lo cation of Indimrator
3-7
-------
estimated in each cell for each pollutant is described by:
where:
A, = ambient afr concentration of pollutant, in cell /due to emissions from all
. facilities imi>acting that cell Gtg/m3),
i = index for grid cells,
j = index for pollutants,
P = index for incinerator facilities,
N = number of Mcmerator facilities modeled,
D* = dispersion ratio for cell / from facility p fog/m3 per g/sec), and
*5p - emission rate* for pollutant./ from facility /> (g/sec).
When a particular grid cell / falls outside the range of a particular facility/,, D¥ is set to zero.
3,1.4 Estimating Human Exposure and Risk
fall withJeth^Hr'^iiX 4-rT ** ^ ^^ ED/BG obse™tions ftnm MARF typically
mDu^o?t^ « r coire?X)IldillS POP^tiom ate summed to calculate thl toJ
mDuot « r o
!T££ ?- ^1??' i03868 Where n° observation ft«m MARF falls within a large grid
cell that cell » Ignored for calculations of aggregate risk. For smaUer 200 x 200 m SfcS?
«M^a?^a=££M^^^
os
on the total density of human populations within the 4 x 4 km boundary of the
We combine aU populations whose ED/BG centroid fall within the lard ceU
mcmerator facility to onhtol he total population for the large
a, ndf r e sma
around the facility. In rare cases where no centroid falls within the large cell at the centefof
the gnd system, we conservatively calculate a density from large gridlells within ran^ ^f Ito
3-8
-------
plant. This average population is then assigned to the large cell containing the facility. As :: |
before, this population is divided equally among the small cells. This adjustment results in a v~ ^
slight over-counting of exposed populations, but its effect on aggregate risk estimates is believed
to be insignificant. \
We generate an estimate! of human exposure for each cell by combining estimated
concentrations of a pollutant with assumptions about daily inhalation volume, body weight, and
background intake of the pollutant from other sources. As mentioned earlier, we assign
populations from the 1980 Census to each of the 840 grid cells for which dispersion ratios are
estimated near each incinerator. For carcinogens, only incremental cancer risks from sludge
incineration are of concern in this study; therefore, background pollutant levels are not
considered in assessing carcinogenic risk. However, background intakes for non-carcinogenic
pollutants are important for estimating non-carcinogenic risks, since background levels contribute
to total exposure, the exposure measure of interest when evaluating the possibility of exceeding
Ojeac3i4XjUutanti:Leach-jcei^ •
BW
where:
"*.'•«.
= exposure to pollutant j for individuals living in grid cell i (mg/kg-day),
Ay = estimated average concentration of pollutant j in grid cell i, including
contributions from all relevant incinerators and excluding contributions
from other sources (/tg/m3);
I. = inhalation volume (m'/day),
10"3 = constant to convert units from /ug to mg,
BIj = background intake for pollutant/, set to zero for carcinogens (mg/kg-day) ,
and
BW = average body weight (kg).
As can be seen from the equation, we conservatively assume that each person residing in a given
grid cell inhales air at the estimated (outdoor) concentration for 24 hours per day for his or her
entire lifetime. We also assume that all of the inhaled pollutant is absorbed into the body, and
thus exposure is effectively equivalent to dose _________
The final step in estimating human health risk is to combine estimates of exposure and
population with pollutant-specific dose-response relationships. For carcinogenic compounds, the
risk to an exposed individual is expressed as the probability the individual will develop cancer
within his or her lifetime as a result of the incineration of sewage sludge. For an exposed
individual hi cell /, we calculate the incremental risk of cancer as:
where:
3-9
-------
Oi - individual cancer risk for individual in ceU / (probability of developbie
cancer from a lifetime exposure), s
M = number of pollutants, and ; :
9i* = human cancer potency for pollutanty (mg/kg-day)'1-
The HEI is defined as the individual for whom this sum is highest:
= MAX CI,
i
where:
SeteT individual "^ (mcremental risk of developing cancer within
N
CP =
where:
CP = across ^ ^^ (e^q>ected cases per
POPi - human population Uving in ceU / (persons),
A-J> - lifespan of average individual (yr), and
N = number of cells.
ss-j-ffis^ssss:
k
-i' ua. a /
exceeding the reference concentration for a given pollutant:
Py = POP.
Pr= 0
and:
where:
3-10
-------
Pij = number of Arsons in cell i exceeding RfD for pollutant j,
RfD, = risk reference.dose for pollutant j, and
NCP, = number of i>eople exceeding the RfD for pollutant j.
i
3.1.5 Estimating Benefits from Regulation
To estimate benefits from the regulation, we first estimate risks under baseline (pre-
regulatory) conditions. We next determine which pollution control devices will be installed at
particular POTWs. We then estimate risks under the new conditions, with estimated emissions
adjusted as appropriate for the presence of additional controls. However, the decision to install
additional controls is influenced by the concentrations of metals in the sludge incinerated at a
particular facility, and unfortunately, data describing those concentrations ate not available for
this analysis except for those 23 incinerators included in the analytic survey of the NSSS. To
model behavior for the larger set of 172 POTWs practicing incineration would require data for
the concentrations of metals in the remaining 149 facilities. In the absence of such data, we use
the following six steps to estimate risks after installation of additional pollution controls and
subsequently to estimate benefits liom the regulation::
(1) Estimate exposure and risk under current conditions for the 23 incinerating
POTWs in the analytic component of the NSSS, using known quantities of sludge
and concentrations of metals for each facility, together with estimated emissions
of organic pollutants. ,
(2) Estimate exposure and risk under baseline conditions for the larger set of 172
known incinerators, after assigning average concentrations of metals to the
sludges burned in these incinerators based on size category, and after scaling
estimates of sludge volumes (from U.S. EPA, 1989g) so that die total sludge
volume matches the total estimated from the NSSS after scaling with sample
weights.
(3) Calculate the ratio of estimates derived in Step (2) and Step (1) to yield a scaling
factor that includes the effects of the increased volume of sludge incinerated,
meteorology, population densities, types of furnaces and controls, and overlapping
plumes from incinerators not included in the analytical component of the NSSS.
(4) Estimate exposure and risk from the 23 incinerating POTWs in the analytic
component of the NSSS, after installation of pollution control devices in response
to regulatory controls. ~. ._ ,..-
(5) Scale the estimate derived hi Step (4) by the ratio derived in Step (3) to estimate
expected aggregate risks after the regulation.
(6) Subtract results of Step (5) from results of Step (2) to estimate the reduction in
health risks to be achieved by the regulation.
3-11
-------
These six steps provide estimates of aggregate risks both before and after implementation
of regulatory controls. Estimates are based on sludge concentrations and volumes from those
POTWs in the analytic component: of the NSSS, and have been scaled to accommodate both the
SUrV6y> "^ ^ aVaikble *w*edfic ^ for ^inerators in the larger
3.2 DATA SOURCES AND MODEL INPUTS
Mathematical modeling of exposure and risk from the incineration of municipal sewage
sludge requires data describing characteristics of the incinerators, the quantity and quality of
sludge incinerated, local meteorology, the locations of human populations, and dose-response
functions for pollutants of concern. This section provides the sources we have used to obtain
these data, and provides tables for key input parameters used for the models.
3.2.1 National Sewage Sludge Survey
Our only source of data for current concentrations of contaminants in the sludee of
individual POTWs is the National Sewage Sludge Survey (NSSS). The survey's analytic
component includes a sampling of sludge from each POTW in a stratified sample of 25 facilities
reporting the use of incineration as a sludge disposal method, and includes measurements of the
concentrabon of 10 metals and over 400 organic analytes. The final analytic data set includes
only 23 facilities, because sludge ifrom two of the original 25 POTWs is incinerated off-site
Of particular interest for this analysis are the measured concentrations of metals; concentrations
of organic contaminants are not known to correlate with emissions from incinerator stacks.
A vTo£0r,63 POTWs responding to the questionnaire (including those in the analytic survey)
the NSSS also provides data for the quantity of sludge incinerated and certain characteristics of
the individual incinerator facilities, including the types of furnace and pollution control devices
in use.
3.2.2 Other Sources of Facility-Specific Data
Additional site-specific data axe available from an earlier analysis of aggregate risks from
the use and disposal of municipal sewage sludge (U.S. EPA, 1989g). That analysis used data
from the 1986 NEEDS Survey and other sources to identify 169 POTWs using incineration to
manage their sludge. Names, facility identification numbers, and latitude and longitude for each
POTW were collected, along with data describing the physical characteristics of their incinerator
facilities (stack height, etc.). For tliis analysis, we have supplemented results from that earlier
study with data collected through the NSSS, to generate a revised data base that includes
information for 172 POTWs using iincineration. For those POTWs not included in the NSSS
earlier estimates of the quantities of sludge incinerated have been scaled so that the total quantity
of sludge incinerated by the larger set of 172 facilities matches the current estimate of the
national total (823,005 DMT/yr) based on data and sample weights from the NSSS The total
3-12
-------
estimated mass of sludge incinerated, in each state, together with each state's contribution to the
set of 172 facilities, is presented in Table 3-1.
3.2.3 Furnace Types
At least three types of furnaces are used to incinerate sewage sludge: multiple hearth,
fluidized bed and electric. Fluidized bed furnaces bum relatively cleanly and tend to operate
at a constant temperature, whereas multiple hearth corabustors are subject to wider fluctuations.
The higher furnace temperatures typical of the fluidized bed combustors tend to improve
destruction of organic pollutants, but can also increase the fraction of metals entering the flue.
Because applicable data are not available to describe emissions from electric furnaces, we
assume for this analysis that emissions from electric furnaces are comparable to those from
fluidized bed combustors.
3.2.4 Air Pollution Control Devices
Pollution control devices can be roughly categorized by the mechanism they use to
capture pollutants, and by their use of wet or dry removal equipment. The four most common
methods of pollution control include: wet scrubbers, dry cyclones and fabric filters, wet
electrostatic precipitators, and afterburners.
Wet Scrubbers
Most sludge incinerators in the U.S. use some type of wet scrubbing device to control
paniculate emissions. Incinerators! tend to be located near POTWs, which provide effluent as
a readily available and inexpensive supply of feed water for the scrubber. Wet scrubbing
devices have demonstrated a long history of success in meeting emission control standards for
paniculate matter.
A typical example is the venturi scrubber. Water introduced at the throat of the venturi
is dispersed by the high velocity g,as stream. Pollutants are dissolved or adhere to the drops,
which are generally larger than particulates and easier to remove. The efficiency of a venturi
scrubber depends on the pressure drop across the throat and the temperature alter the venturi.
If the temperature is too high, further evaporation occurs and condensation is reduced. Cyclones
may be used upstream of the scrublber for better removal of particles with diameters larger than
10 micrometers. The gas stream can then be bubbled through impingement trays filled with
water (located downstream of the scrubbers) which further entrain solid particles and increase
the removal of pollutants from flue gas.
3-13
-------
Table 3-1
Mass of Sludge Incinerated and Number of Incinerators by State
Alaska
California
Connecticut
Florida
Georgia
Hawaii
Indiana
Iowa
Kansas
Louisiana
Maryland
Massachusetts
Michigan
Minnesota
Missouri
Nebraska
Nevada
New Hampshire
New Jersey
New York
North Carolina
Ohio
Oregon
Pennsylvania
Rhode Island
South Carolina
Tennessee
Virginia
Washington
West Virginia
Wisconsin
TOTAL
•rWoriluo It iwrn;
Number of
Facilities
in Analytic
Survey*
1
1
1
2
1
1
1
2
4
1
1
1
.3
2
1
23
Quantity
Incinerated
in Analytic
Survey
(DMT/year)*
11,748
5,297
4,137
182
2,786
277
11,500
5,705
152,012
59,874
3,975
2,268
66,998
22,731
6,622
356,110
Percent of
Total !
Quantity
for Analytic •
Survey*
3.30%
1.49$
1.16%
0.05%
~*
0.78%
0.08%
3.23%
1.60%
42.69% '
16.811%
1.12%
0.64%
18.81%
6.38% :
1.86%
100.00%
Number of
Facilities
in Larger
Sample*
2
9
11
1
6
2
1
3
4
4
. .3,. ... ...
7
14
3
4
1
1
2
9
26
3
12
1
16
1
2
5
11
-4 .
1
,3 .-. - .
172
Quantity
Incinerated
in Larger
Sample*-"
(DMT/year)
446
40,848
19,381
4,137
6,040
3,914
11,018
13,004
3,124
10,952
•™ 13,536
22,020
174,858
68,619
44,640
4,368
114
2,511
12,640
105,023
3,717
139,504
2,060
46,333
3,308
4,917
10,002
42,930
4,233
529
4,280
823,005
Percent
of Total
Quantity
for Larger
Sample*"
0.05%
4.96%
2.35%
0.50%
0.73%
0.48%
1.34%
1.58%
0.38%
1,33%
1.64%
.2.68%
21.25%
8.34%
5.42%
0.53%
0.01%
0.31%
1.54%
12.76%
0.45%
16.95%
0.25%
5.63%
0.40%
0.60%
1.22%
5.22%
0.51%
0.06%
0.52%
100.00%
... , . .. - - •-* Quantities shown have not been scaled by samole
weights; scaled total is 823,005 dry metric tons per year.
"Based on list of 172 incinerating POTWs compiled from U.S. EPA (1989g) and NSSS
iSSSnl f ^ f^TPcla±n0t iDCluded * *" NSSS *"" been estimatcd !">* '"*»» "Toted * *e
1986 NEEDS Survey (U.S EPA, 1989g). For this analysis, estimates for those excluded plants have been
proportionately scaled so.that the total quantity of sludge for all facilities matches the total estimated with
sample weights from analytic survey of NSSS. ""»«=" wua
3-14
-------
Dry Cyclones and Fabric Filter; :
« '
i
Dry removal devices rely on inertial, adhesive or electrical forces. A dry cyclone is
typical of an inertial force separator. In the vortex chamber of a cyclone, the gas is set in a
rotational pattern so that particles and droplets are separated from the gas by centrifugal action.
This process is most effective for the more massive particles and droplets.
A bag, or fabric filter, uses adhesive forces to remove pollutants. Fabric filters use a
porous fabric that traps particulates as the gases pass through a series of suspended fabric filter
tubes. The collected particulates are periodically shaken from the fabric for accumulation and
removal from the bottom of the apparatus. If lime is added to the gases the acids are
neutralized, and the fabric filter collects the lime and other particles before the gases leave the
stack. The first sewage sludge incinerator with a dry scrubber and fabric filter system has been
designed for a POTW in California and should soon be operating.
Wet Electrostatic Prccipitators
Wet Electrostatic Precipitators (ESPs) operate by electrically charging particles in the
gases as they flow through a chamber. Oppositely charged plates or tubes then attract and
collect the particles. The particles are periodically washed off the collection plates and gathered
by an ash slurry system. Wet ESPs have been specifically tested as a retrofit option for sludge
incinerators. Particles and droplets with sizes as low as approximately 0. 1 pm can be collected
using a wet ESP (Brauer and Varma, 1981). However, small particles and drops tend to migrate
more slowly than larger ones. To capture the lower size ranges, the collection areas must be
large and gas velocities low. Of special importance is the effect of temperature on collection
efficiency. The electrical resistivity of the particles or droplets determines how readily they
migrate to the collection surfaces. Graphs of electrical resistivity against temperature from solid
waste incinerator samples show a very pronounced peak occurring between 100 and 200 °C
(Vancil and White, 1988). Away from the peak resistivity temperature, resistivity fails off
rapidly. Moreover, above a resistivity of about 1010 ohm-cm, a "back corona" develops, which
can significantly reduce migration. Both of these factors affect the optimum temperature for
operation of a wet ESP. Results of tests on sludge incinerators show that removal efficiency
decreases dramatically as temperatures depart from design temperature.
According to Gerstle and Albrinck (1982), approximate ranges of particle sizes targeted
by the pollution control devices just discussed are: .....
Wet Scrubber 0.05 to 300 jaa
Filter 0.05 to 16
Cyclone 4 to 800 /«n
Electrostatic Precipitator 0.05 to 12
Particles of fly ash range in size from 1-100 /*m.
3-15
-------
Afterburners
oxidize, in a secondary combustion chamber organic compounds not
primary incineration orocess Th»» cnXon> ,,~~,u.,^~_ j i_. _- .
A ~* J j • w~«~~,«i « sctonoary comousuon chamber organic comoound* nn*
±S± 1-5 frT7 facte^!«««. The secondary combSn j£3S2fLS
gaseous or liquid ruels or by exposing the primary combustion gases to a bed of cataly^clgente
A review of twelve studies shows that afterburners remove up to 98% of most tra •
poUutants and 99% of dioxins and rurans (Environment Canada, 1986). ^ Ot&m*
3.2.5 Current Inventory of furnaces and Pollution Controls
Table 3-2 summarizes our assumptions for the types of furnaces and poUution control
devices operating at each of the 23 incineration facilities Muded in the Sy^ sfrvey^e
fl -/T *A^A^ SCen ft°m ^ ^^ only tnree °* *^e. ^3 feciiities modeled \
tiuidiz^ bed rurnaces, and aU incinerators except St. Paul's rely on conventional wet
to control emissions of metals. Only one of the 23 facilities (in Martinez CaSS
to operate an afterburner currently. Similarly, of the 172 plants in our'
is the only facUity known to operate an afterburner.
3.2.6 Expected Response to Regulatory Controls
(NSP^ i . °f ^ Cl6an ^ Act' New Source Performance Standards
(NSPS) require all new sludge incinerators to install scrubbers meeting the National AmbiemAir
Quality Standard (NAAQS) for peculate matter. Data gatheL foTindfviS ?ll±
incinerators show that most facilities', already have an air pollution control device (i e^crubber
inSd^^6 ^ ^ CT!?ustion V™** ^d the type of poUution control
installed will affect poUutsmt emissions. For this analysis, we consider onlv th^
6 ? *^ «** d^ices f<* «*«*, stack emissions.' SLSTSdS £
^m?!r^^^yi^flrtii^rf^
to which a highly exposed individusd might be exposed. These maximum concentration^ are
C
' ceswso
result in unacceptably high concentrations wiU be required to install additional poUution control
devices. For ^simplicity, we consider only one additional control technology. To reduce meS
STV faahtyJ2uld «**; wet electrostatic precipitator (ESP) operating at either high or
medium efficiency; reduction of organic emissions is addressed by adding an afterburner Based
on a standard of 100 parts per million of total hydrocarbons (THC), EPA predicts that existing
incinerators will be able to comply with crtieria for emissions of organic poUutants with minimal
changes to current practices. Increased use of afterburners is not expected as a result of the
regulation. Table 3-2 shows the poUution control devices expected to be instaUed by the 23
incinerators in the analytic survey. aumcu uy uic zj
3-16
-------
1
!
u
s
ffi a
S J
I
I
i§
u u u
III
I
DO
s
1
S H 8
*2l?S
«
I
I
Ot
a
I
1
S
frl
a
I
•a
1
-------
XT A Ar 5dditional consdaim is. that ambient concentrations of lead must meet the lead
NAAQS of 1.5 /tg/m3. According to the modeling performed for this analysis this
concentration is not exceeded by any of the 172 incinerators considered. In fact, the highest
concentration simulated under basieline conditions is about 0.14 ug/m3, or just under 10 percent
of the NAAQS standard.
3.2.7 Data for Estimating Emissions of Inorganic Pollutants
Incinerator emissions depend on the type of furnace operating, the quantity of sludge
incinerated and the effectiveness of pollution control devices. For inorganic pollutants, the
quantity of sludge incinerated and its concentratioas of contaminants determine the mass of
pollutants potentially available for emissions. Final emissions are determined by the combustion
process and the pollution control devices installed. Characterizing emissions from a variety of
combustor and pollution control combinations therefore requires estimates of the rate at which
each metal is fed into the incinerator, the metal-specific removal efficiencies of the combustor,
and the removal efficiencies for the pollution control devices.
Sludge incinerated at the 23 facilities in the analytic survey of the NSSS was tested for
ten metals; results are summarized in Table 3-3. To estimate concentrations for the remaining
149 plants in our larger sample, POTWs in the analytic data set have been divided into four
categories based on the volume of wastewater treated per year. Average concentrations of each
metal have been determined for the plants in each of these four categories and assigned to
comparable size plants not include*! in the analytic survey. Results are listed in Table 3-4.
To predict emissions of metals for this analysis, we use the average concentrations of
metals described in Tables 3-3 and 3-4, together with estimated feed rates for each facility, to
predict the mass of each metal entering the incinerator per unit time. We then adjust this mass
for the "removal efficiency" achieved by each type of incinerator and control device to predict
emissions. Removal efficiencies liave been calculated by averaging available test results from
incinerator facilities withsimilar furnaces and control devices.
One important consideration in estimating emissions is that different types of pollution
controls target different particle amd droplet sizes for optimum collection. The combustion
conditions occurring in different furnace types will produce different arrays of particle and
droplet sizes. For-example, increased air velocities through the combustion bed will increase
particle entrainment. Changes in the amount of excess oxygen available will affect the
vaporization of volatile metals (U.S. EPA, 1987h). As a result, removal efficiencies depend on
particular couplings of furnace and control.
Furnace temperature is also important in determining metal emissions at a given furnace.
A study by Gerstle and Albrinck (1982) examined the volatility of metals commonly found in
sewage sludge, and compiled data to investigate the relation between furnace temperature and
percent of input metals that are emitted to the atmosphere. At conventional temperatures for
incineration (760-815 °C) arsenic, icadmium, mercury and zinc are volatile. Lead is potentially
volatile at temperatures above 980 °C. These metals tend to be emitted in a vapor state while
3-18
-------
eo
.9
I
£
£
•8
I
1
f
a
3
I
I.
^
I
I 9
It
OQ
O -j V)
m S3 r-
tn " «n vi ii
5 8 £
I""" ^^ ^Q
O; o\ V) —«
2 w ™ ° « <-»
N S 2
000^35
S
10
SI
2
«•> Si "*
en jo in
ooooooooooooooooooooooo
£ *!
^ «
.c
II
1s
Z n
If
*3 §
u. P-
I
X
I
I1
^ 55
S
1
if
x «
Hill
ON
a
53
-------
Table 3-4
Metal Concentrations by Flow Group
For Sludge Incinerated at Plants in the Analytic Survey
Average Average
Concentration Concentration
for POTWS for POTWS with
Arsenic
Cadmium
Chromium'
Copper
Lead
Mercury
Molybdenum
Nickelb
Selenium
Zinc
with 0-1 MOD
(mg/kg)
6.64
120.7
t
651.7
916.5
207.1
2.22
9.34
191.25
10.21
1941
1-10 MOD
(mg/kg)
7.4
58.38
211.69
621.3
259.8
2.0
9.04
65.84
4.16
1627
Average Average
Concentation Concentration
. for POTWS with for POTWS with
10-100 MGD
(mg/kg)
3.088
i 6.388
150.675
346.1
174.4
• 2.863
, 3.675
53.212
2.238
1670
100+ MGD
(mg/kg)
5.2
3.7
26.5
954
55.7
3.7
0
: 15,9
4.100
460
"As total chromium.
bAs total nickel.
3-20
-------
others (including chromium, copper and nickel) are generally emitted as fly ash. Graphs of
temperature against metals emitti^d- for both controlled and uncontrolled emissions show an
increase in the concentration of volatile metals in stack gas with higher furnace temperature.
Conventional wet scrubbers appeair relatively ineffective for capturing volatilized metals.
Mercury is the most volatile of the metals considered. Both inorganic and organic
mercury compounds can volatilize at temperatures as low as 365°C (Dewling et al, 1980). As
a result, mercury is emitted as a vapor or as volatilized mercury compounds. A study conducted
in 1980 at the Northwest Bergen Sewer Authority recorded concentrations of metals in the feed
sludge, bottom ash, scrubber water and stack emissions. According to these data, which were
normalized for the mass balance, 98% of the mercury in the feed sludge was emitted to the
atmosphere. Scrubber water contained about 2% and only 0.4% remained in the ash. The
incinerator used a fluidized bed combustor with a wet scrubber and average furnace temperature
of 788 °C. Discussions from other sources agree that typical pollution control devices are not
effective in removing mercury from incinerator emissions.
Chromium is most likely to be emitted from incinerators in either a hexavalent or
trivalent state. Hexavalent chromium is a suspected human carcinogen, whereas trivalent
chromium is a necessary trace element for humans. In general, the hexavalent chromium is
expected to be more common in emissions from furnaces operating at higher temperatures with
more available oxygen. Based on results from analysis of chromium emitted from Site 8, we
assume that 1.3 percent of that emitted from a fluidized bed incinerator is hexavalent. For
multiple hearth furnaces, testing at multiple sites has shown that hexavalent chromium ranges
from about 1 percent of total to about 13.2 percent of total emitted chromium; the value of 13.2
percent is assumed since it represents a conservative value that possibly over-estimates cancer
risks. For nickel, the carcinogenic species is nickel subsulfide (NiaSj). Based on test results,
we assume that one percent of the! nickel emitted from the average multiple hearth or fluidized
bed incinerator is in this form.
Ideally, efficiencies for each type of furnace and control device would be evaluated
individually. For example, the removal efficiency of a furnace would determine how much of
the metal contaminant in the feed sludge would remain in the bottom ash, and how much would
be emitted as uncontrolled emissions. A scrubber efficiency would relate the quantity of
contaminant entering the scrubber (uncontrolled emissions from the furnace) to the quantity
leaving the scrubber and entering the atmosphere. Emissions from a particular incinerator could
then be determined as the product of the efficiencies of its components (e.g., furnace and
scrubber). However, available data are not always sufficient for accurate determination of the
efficiency of individual components in the incineration process. In some cases, efficiencies can
only be estimated for selected couplings of furnaces and controls.1
'As mentioned above, removal efficiencies are sensitive to the operating conditions of
furnaces and pollution control devices. These variables are not modeled explicitly in this
analysis, although they are captured implicitly if conditions at those plants sampled are
representative of the larger inventory of incinerators.
3-21
-------
Sampling incinerator stack emissions and analyzing the samples is a complex, expensive
process. The type of material combusted will to some extent affect the form of the metals
entering the flue and therefore the effectiveness of the controls. For this reason data from *
incinerators burning materials other than sludge (e.g., municipal solid waste incinerators) cannot
be used to represent the combustion and emission processes in a typical incinerator for sewage
sludge. The availability of useful data is further reduced if one discards poor quality data with
likely errors in sampling procedures. Where removal efficiency data, are reported or can be
calculated, they are usually based on the ratio of the rate at which the pollutant is emitted to the
inflow rate for contaminant mass.
For predicting emissions udder baseline conditions, three separate estimates of average
removal efficiencies have been prepared for three furnace/control couplings: for multiple hearth
incinerators with wet scrubbers, for fluidized bed incinerators with wet scrubbers, and for
multiple hearth incinerators with dty cyclones and fabric filters. Table 3-5 contains results from
tests performed on multiple hearth incinerators with wet scrubbers. Data from eight such
incinerators were used; data available from two other sites, the Osborne Wastewater Treatment
Pknt in Greensboro, NC, and Site 9, were not used. For the Osbome plant, estimated removal
efficiencies were greater than 100 percent, and for Site 9 they were significantly lower than for
the other eight sites and were therefore discarded as outliers. Available estimates for each metal
have been averaged to yield the values in the last column of the table. For this analysis we
assume that these average removail efficiencies are achieved by the 20 plants in the analytic
survey (and the 164 plants in the Lirger sample of 172 facilities) using multiple hearth furnaces
with wet scrubbers.
Table 3-6 contains estimates for fluidized bed furnaces with wet ESPs. Data from Glens
Falls Waste Water Treatment Plant, Glens Falls, New York were excluded because the inflow
rates were very low. As with multiple hearth incinerators, we have averaged these values to
derive the results listed in the last column of the table. These averaged removal efficiencies are
applied to the 3 facilities in the analytic survey of the NSSS (and 8 facilities in the larger
sample) known to use fluidized bed'fumaces with wet scrubbers. Table 3-7 summarizes removal
efficiencies for wet scrubbers from both multiple hearth arid fluidized bed furnaces.
In response to the regulation, some (but not all) POTWs are expected to retrofit their
incinerators with additional pollution control devices. As mentioned above, further removal of
metals is assumed to be achieved by installation of wet electrostatic precipitators. We assume
that wet ESPs remove 95 percent of all remaining metals when operated at high efficiency, or
90 percent when operated at medium efficiency. An important distinction should be noted
between efficiency data for wet ESPs and for all other types of controls for metals. The 90 and
95 percent removal efficiencies assumed for wet ESPs apply only to the control device itself-
control efficiencies reported in Tables 3-5 through 3-7 pertain to specific couplings of furnace
types and control devices. Data for the wet ESP are therefore useful for evaluating this
technology as a retrofit option and removal efficiencies can be applied to the concentrations of
metals emitted from the control devices already installed. For comparison with values listed in
Tables 3-5 through 3-7 and used in this analysis, Table 3-8 summarizes results from other
studies that investigated emissions of metals from incineration of either sludge or municipal solid
wastes.
3-22
-------
I
00
ON
O
en
£
en
en
oo
oo
s
ON
ON
IT> ON
<3S - «-'
ON ON
en ON ON en '
v> ON
ON ON
ON
S
ON
ON
ON
ON
ON
ON
ON
ON
Xfa
&
oo
4
£
00
£
00
€
00
00
a
ON
ON
ON
ON
gj
O
oo
oo
ON
oo
ON
ON
ON
O
ON
ON
ON
o
oo
O
ON
ON
ON
ON
ON
O
OO
S 8 !
ON
ON
ON <*\
ON CJN
ON
ON
&
' 8
1 5?
ON
ON
en
8
en
OO
ON
en
oo
SN
o
ON
ON
ON
£
S3
CM
en
ON
ON
8 Pi *
ON
00
ON
I s
_o .2 .s
ii>i
i- a) ea
< 03 O
1 £ -a i 1 1 a
e a 3 I I 2 s
-------
S £«•• 55 % &
O\ ON ^ ON ON
ON ON ON ON ON
e
.O
I
CO
ON ON
ON ON
1
E
t-l
.§
o
1
ft
"8
•i
I
ON ON
8 8 °^ ON
vo
ON ON
ON ON
jr* X «S <»
°i S *~! <-. •
ON g •-- — '
ON g
S
§1
en
-------
Table 3-7
Summary of Removal Efficiencies for Metals
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Nickel
Selenium
Zinc
Multiple Hearth With
Wet Scrubber*
97.48 !
97.33
88.54
99.03
99.93
91.59
98.98
99.80
99.90
Fluidized Bed With Wet
Scrubber1'
99.91
99.997
99.27
99.91
99.98
99.89
99.84
98.53
99.46
•From Table 3-5.
bFrom Table 3-6.
3-25
-------
& I
5
1
a;
S
«
&
1
CO
.2 .§
= .&
co
a-,
9,
u
•s
SS
U
8 3
«:&
& I
Q£
1
55
*3
,
u g
I 8
i'lj
*5l
lit
8,
K
rfl
vo
^
en
31
(C ^
B ' I
* JJ*
311
£ §£
S
-------
E
CO
i °
I 1
^ a
s §
•s
1
I
H
Air Pol
I
•" o
0-9
vq «o
??
^ o\
tfc
o\
oo"
3 o-
*-i T-H
V A
1
S"
i
I
d
3: "O ^3
S St2
•ill
8 >,
f
53
S
a
4_t i^ ^3
.S -a E
oe
•a
i
r- oo
CT\ <>\ cS cy\ os os
1 1 1 1 1
Siz
3
"S,
I
l
&
1
-------
CO
-------
?
I
4d-
oo
m
I
I
o
S
a
1
u
'o
S
W
•a
>
0
I
•W
a
<3
I
•J3
.2
•a
1
S
<4-<
O
8,
£
>.
•a
S
03
S
o
00
1
I
•o
2
§
1
a
.2
.&
§
Q
"8
a
S
^
' ' i i i i t
^5
?-N S?
u^ ,*•, ***?*
— j*i» ^3^
g f^l |
,}. t^gtiS' Jn"
00 ^^ OO ^^ £|
ON i ^ ON en t ON
Ml
lllllli
r §
f jj
s ^1 .
> g J §
ill*
Sltfl
sHI
§ 5 5 vi
S> tn 8 a
&
1
en
UJ
•a
.5*
O
S
•*
j)
iJ5
1
1' '
1
H
S>
"O
3
55
o
r
$
a
§t^ \O oo *• ^ ^y ^f ^>( ^^
*TJ ^^ tn ^o ^T *o ^ \o
00 OO f"* €fi *"• CS O\ ^** f«4
i i t i i o\ O\ oo v\ Ox CT\ ON ON i \o
.
• !
4 * !
tft W^ tf ' M% WS
-« O (SI PI (S
00 liilt tittt
I :
o 3 -a o a -3 o 3 -a
Hill ||||| Hill
| e !
1 1
1 1 ^
>• p^ tQ
- ?r ^
i
&
1
C/3
1
•S3
S
t. •
•+* "S
E -| .
•g> .-
J3 -
^'l
ii
!&•
58
*1 ;
^ tt]
JS-a
flfl «
it -
£&
*'- ' -
ii i i
o\ oo vi o\
S S; S g
II II
•2 «2 ^ "
' s 5S si "a
o o o o
U) QO'
O "w? ^? "OT
"S. S ^ 1 § a a
S ? 8 •§ -i § §
s S? * 3 iJ <-. «
S | § E |> J §
• ° o o-. oS 2_.
& 3 & S JP 4) $1}
fel b« 8s i -c«
rf s -9 -a wej-se
|| ||0 £fi |||
^s ? •§ g o« > -I ^
s-ai f?E3 fi'fi ^E<2
g,
•f
53
]3
S
a
o
|
8
g
1
?§
«-a
a-?
O*J {J
o o
1-s5
S —
eu 55
R
CM
o
1
en
3
-------
I
I
-------
3.2.8 Emissions of Organic Pollutants
«
To date, no useful correlation has been identified between the mix of organic compounds
entering a sludge incinerator'and the mix of organic compounds released in stack emissions. For
this analysis, emissions of organic pollutants are considered a function of sludge feed rate and
the type of furnace, and are not related to the concentrations of organic compounds in the sludge
feed. Assumed profiles of emissions of organic compounds are based on measurements obtained
from seven actual incinerators for wastewater sludge. Of these seven facilities, five were
multiple hearth incinerators and two were fluidized bed combustors. This distinction is
important because the improved burning conditions in fluidized bed combustors allow more
complete oxidation of organic compounds and minimise the creation of products of incomplete
combustion (PICs). !
For both our "best estimate" and "worst case" estimate of organic emissions, we
considered all compounds tested for in at least one facility sampled in the monitoring studies,
regardless of whether the compounds were detected. For the "best" estimate, expected emissions
are based on the arithmetic mean of values for the multiple samples taken for each furnace type.
Samples with concentrations below limits of detection are assigned values corresponding to the
appropriate detection limit. Where possible, each non-detect observation has been assigned the
detection limit concentration for the particular site and chemical tested; otherwise, it has been
assigned an average concentration based on detection limits at other sites or for similar
compounds. Results are listed in the first and third columns of Table 3-9. Based on the
available sample of measured (or assigned) values, a 99th percentile confidence or "worst case"
estimate has been derived for the average emission rate for each pollutant for multiple hearth
or fluidized bed combustors, as listed in the second and fourth columns of Table 3-9.
These "worst case" estimates for emissions are further adjusted to account for additional
organic pollutants not tested. The Municipal Waste Combustion Study Report to Congress (U.S.
EPA, 1987a) asserts that "a significant portion (80% or more) of the organic emissions from the
stacks of municipal waste combustors have not been identified and quantified. Although some
portion of the mixture may be carcinogenic, the carcinogenic fraction, its composition, and its
potency remain unknown." If a similar fraction of emissions from sludge incinerators has not
been identified or quantified, calculations based on known emissions might understate true
exposure and risk. To adjust for this possible error, we assume that the average cancer potency
of these unknown compounds is comparable to that for the chemicals evaluated. Assuming that
in fact 80% of organic emissions have not been characterized, estimated health risks from
incineration of wastewater sludge are increased by a factor of five to derive a conservative
estimate of total risk.
3.2.9 Population Data
As discussed in Section 3.1.4, we obtain population data from the 1980 Census at the
enumeration district or block group (ED/BG) level of aggregation, as provided within PC-
GEMS. The MARF data base of PC-GEMS provides a convenient listing of population .by the
latitude and longitude of each ED/BG centroid.
3-31
-------
Table 3-9
«
Unit Emissions of Organic Pollutants
-t(g/s «imitted per kg/s sludge feed)
1 ,2-dichloroethane
1 , 2-dichloro benzene
1 ,2,3 ,4,6,7,8-hcptachlorodibenzofuran
1,2,3,4,6, 7,8-heptachlorodibenzo-p-dioxin
] ,2,3,4,7,8-hexachlorodibenzo-p-dioxin
1 ,2,3 ,4,7,8-hexachlorodibenzofuran
1 ,2,3 ,4,7,8,9-heptachlorodibenzofuran
1,2,3,6,7,8-hexachlorodibenzo-p-dioxin
1,2,3,6,7,8-hexachlorodibenzofuran
1 ,2,3 ,7,8-pentachlorodibenzo-p-dioxin
1,2,3,7,8-pentachlorodibenzofuran
1,2,3,7,8,9-hexachloroditxMzofiuan
1 ,2,3,7,8,9-hexachiorodibenzo-p-dioxin
1,2,4-tricfalorobenzene •
2-chlorophenol
2-methylphenol
2,3,4,6,7,8-hexachlorodibenzofuran
2,3 ,4,7,8-pentachlorodibenzofuran
2,3,7 ,8-tetrachlorodibeazofuran
2,3 ,7,8-tetrachlorodibenzo-p-dioxin
1 , 4-dinitrotoluene
2,4-diinethylphenol
2,4-dichlorophenol
2,4-dinitrophenol
2,4,5-trichlorophenol
2 , 4 , 6 -trichlorophenol
Multiple
Hearth
Mean*
8.89x10*
7.36X10-4
2.08x10-*
3.30xlO«
1.15x10^
3.66x10*
1.83xlO-»
1.45x10*
8.77X1O9
1.38xlO»
3.21x10*
1.06x10*
2.41xlO»
7.13xlO3
6.33x10-*
5.22xlOs
2.27xlO*
1.32xl07
1.64xlO7
1.15x10*
9.46xlO5
8.17xlO*
7.39xlO<
5.23xlO3
5.08xlO-5
l.OOxlO3
Multiple
Hearth
99*116*
2.11x10*
1.67xlO3
2.08xl04
9..41X1O1
l.lSxlO9
3.66xiO«
1.83xlO»
2.43xlO»
8.77x10-*
3.19x10*
3.21xlO«
1.06x10*""
2.41x10*
2.21xlO2
2.06xlO-3
5.22xlO-5
2.27x10^
1.32xlO-7
jt.02xlO-l
2.82X1019
9.46xlO"s
2.6?xl03
2.41xlO3
1.58xl02
S.OSxlO'5
3.24x10-'
Fluidized
Bed
Mean"
2.83x10*
9.78x10*
1.51xlO"
1.03x10"
l.SOxlO12
9.70xl01J
3.26xl012
1.70xl012
4.94xlO12
1.91xl012
8.71xl012
1.41xlO12
2.23xl012
1.41xlO5
9.39x10*
4.08xlO«
4.05xlO12
5.29xl012
3.98xlO12
2.06xl012
2.14x10*
1.41x10*
1.88xlO5
3.43X1O4
2.08x10*
2.82xlO5
Fluidized
Bed
99%iled
8.79x10*
3.07X1O5
l.SlxlO11
. 1.03x10"
i.SOxlO12
9.70xlOw
3.26xl012 .
1.70xlO)2
4.94xl012
1.91xl012
8.71xlO12
1.4lxlO12
2.23xl012
4.69X1O5
3.12x10*
4.08x10*
4.05xlO12
5.29xlO12
3.98xlO12
2.06xl012
2.14xlO«
4.69xlOs
6.25xlO-5
1.14xlO3
2i08xlO8
9.40x10-*
(continued next page)
3-32
-------
Table 3-9 (continued)
«
Unit Emissions of Organic Pollutants
(g/s emitted per kg/s sludge feed)
3 ,3 '-dichlorobenzidine
4-methylphenol
Acenaphthene
Acenaphthylene
Acetonitrile
Acrylonitrile
Aldrin
Anthracene
Benzene
Benz(a)anthracene
Benzo(a)pyrene
Benzo(b) fluoranthene
BenzoQc) fluoranthene
Benzoic acid
Bis(2-chloroisopropyl) ether
Bis(2-ethylhexyl)phthalate
Bis(2-chlorethyl) ether
Butylbenzyl phthalate
Carbon tetrachloride
Chlordane
Chlorobenzene
Chloroform
Chrysens
Di-n-butylphthalate
Dibenz(a,h)anthracene
Multiple
Hearth
Mean*
9.59x10^
5.92xlOs
1.64xlOs
2.04X10"5
9.58xlO3
2.19xl02
4.28xlO-3
1.07xlO"5
e.oixio-2
7.56xlO-5
3.29X10-1
2.21xlOs
1.92xlO"5
1.10x10-'
2.55X1O5
1.02xlO-3
4.95xlOs
2.08xia5
3.35xlO-5
3.33X10-1
9.72xl(H
1.65xlO-3
7.91x10^
1.80xlO-s
4.86xl(H
Multiple
Hearth
99«ileb
9.59xlO-s
5.92x^0^
4.13xlO-J
2.04x10^
9.S8X10-3
3.68xlO-2
1.39x10-*
1.07xlO-s
1.74x10-'
7.56x10^
l.OSxlO-3
2.21x10^
1.92xlO-s
1.10x10-'
2.55xlO-5
2.05xlO-3
4.95x10^
2.08xlO-5
9.82xlO-s
3.98X1O4
1.86x10-'
4.38xlO"3
7.91xIO-«
LSOxlO'5
4.86xlO'5
Fluidized
Bed
Mean"
4.31x10^
l.SSxia8
7.58X10*
2.96X108
—
1.61xlO-s
3.75x10^
7.44x10*
2.00xlO-«
1.20x10*
1.41X1O5
2.13xlO«
1.73x10-"
5.00x10*
1.44x10^
1.34x10-'
1.41X104
4.61x10*
3.84x10*
5.63X10-1
2.12x10^
1.94xlO"3
1.46x10-*
l,60xlO-7
4.81X10"8
Fluidized
Bed
99«iled
4.31xlO-«
1.38X10-8
7.58x10^
2.96x10*
_
5.31x10^
3.75xlO-5
7.44x10^
6.64x10-*
1.20x10*
4.68X10-5
2.13x10*
1.73x10*
5.00x10*
1.44x10*
4.42x10"
1.41x10*
4.61x10*
1.26x10^
5.63x10-*
7.03x10*
6.45xlO3
1.46x10*
1.60x10-'
4.81x10*
(continued next page)
3-33
-------
Table 3-9 (continued)
Unili Emissions of Organic Pollutants
\ (g/s emitted per kg/a sludge feed)
Dieldrin
Diethylphthalate
Ethylbenzene
Fluoranthene
Fluorene
Hexachlorobenzene
Hexachlbrobutadiene
HexachloFocyclopentadiene
Hexachloroethane
Indeno (1,2,3-cd) pyrene
Isophorone :
Methyl ethyl ketone
Methylene chloride
N-Nitroso-di-n-propylamine
N-Nitrosodiphenylamine
Naphthalene
Nitrobenzene
Other-pentachlorodibenzo-p-dioxin
Other-tetrachlorodibenzofuran
Other-tetrachlorodibenzo-p-dioxin
Other-pentachlorodibenzofuran
Other-hexachlorodibenzo-p-dioxin
Other-heptachlorodibenzofuran
Other-heptachlorodibenzo-p-dioxin
Other-hexachlorodibenzofuran
Multiple
Hearth
Mean*
3.88x10-'
2.95xlO-s
1.76xl03
4,55x10-"
2.14X1O5
8.53x10*
5.26xlO*
2.10x10-'
1. 88x10-*
5.97xlO-s
1.42X10"5
4.94X10'3
1.43x10'
5.61xlOs
3.12xlO'5
8.36x10-'
2.23xlO's
2.25x10^
7.00xlO-7
1.73xlO-7
5.23xlO-7
2.41xlO"«
1.39xlO-»
2.95x10^
7.07X10"8
Multiple
Hearth
99%ileb
1.25xlO-z
2.95xlO-s
3.22x10-'
4.55x10-"
2.14x10-'
8.53xlO-s
5.26xlO-3
2.10x10^
i.ssxio-1
5.97x10^
1.42x10^
4.94x10-'
2.62x10-'
s.eixio-5
3.12X10-5
2.49xlO-2
2.23xlO-5
5.75x10^
7.00xlO-7
3.31x10-'
5.23xlO-7
3.00x10*
1.39X10-8
8.12x10-*
7.07x10^
Fluidized
Bed
Mean"
5.63xlO-5
1.02xlO-7
9.23x10*
5.68x10*
7.05X109
2.40x10*
3.16x10*
2.55x10*
3.34x10*
2.89x10*
7.82X10-9
4.70xlOs
2.14x10*
1.21xlO*
1.79xlO<
1.40x10*
1.15x10"
1.08x10'°
2.58x10"
7.66x10"
1.17x10"
1.15x10"
1.48xlO"
2.87x10"
Fluidized
Bed
99S6ile'1
5.63xlO«
1.02xl07
2.94xlOs
5.68x10*
7.05x10*
2.40X108
3.16x10*
2.55x10*
3.34xlO*
2.89x10*
7.82x10*
1.03x10^
2.14x10*
1.21xlO*
1. 79x10*
1.40xlO*
1.15x10"
1-OSxlO10
2.58x10'"
7.66x10"
1.17x10"
1.15x10"
1.48x10"
2.87x10"
(continued next page)
3-34
-------
Table 3-9 (continued)
«
Unit Emissions of Organic Pollutants
. (g/s emitted per kg/s sludge feed)
PCBs
Pentachlorophenol
Phenol
Pyrene
Tetrachloroethylene
Toluene
Trans 1 ,2-dichloroethene
Trichloroethylene
Vinyl chloride
Multiple
Hearth
Mean'
3.48x10-*
1.52xlO-3
7.85X10-1
2.41x10^
4.27xlO-3
3.79xlO-2
7.31X10-5
1.02xlO-3
3.73xlO-3
Multiple
Hearth
99%ii^
4.49*10-*
4.67xlO-3
1.76xlO-3
2.41x10-*
1.22xiO-2
1.11x10-'
2.14x10-*
2.03xlO-3
6.45xlO-3
Fluidized
Bed
Mean"
6.11xlO-«
7.52x10-*
9.42x10^
3.78x10^
2.75X10"5
3.38xlO-3
4.86x10*
4.94x10^
7.85xlO<
Fluidized
Bed
99%iled
e.iixio-4
2.50x10-*
3.12X10'3
3.78x10-*
8.69x10^
1.09x10-*
4.86x10^
1.37x10-*
2.60X10-5
•Mean of available emissions data for five multiple hearth furnaces, with non-detect observations set to
detection limit.
b99th percentile confidence limit for estimated mean of available emissions data for five multiple hearth
furnaces, with non-detect values set to detection limit.
"Mean of available emissions data for two fiuidized bed furnaces, with non-detect observations set to
detection limit.
^99^ percentile confidence limit for esti mated mean of available emissions data for two fiuidized bed
furnaces, with non-detect values set to detection limit.
3-35
-------
3.2.10 Meteorological Data
PC-GEMS also includes a Stability Array (STAR) Idatabase. Average wind velocity
wind directions, mixing heights and stability categories from STAR stations (usually airports)
across the country are compiled to provide a national meteorological database. As implemented
in PC-GEMS, these data are accessed automatically; we use the STAR station nearest to each
incinerator for describing average weather conditions at the site.
3.2.11 Health Effects Data
As discussed in Chapter 1, health effects are grouped into two categories: cancer and
non-cancer effects. For carcinogenic pollutants, a human cancer potency slope is characterized
and no threshold in exposure is assumed to exist. For most non-carcinogens, a threshold dose
has been identified by the U.S. EPA below which no adverse health effects are expected These
threshold doses are termed "Risk Reference Doses" or RfDs. For this analysis, the primary
source for our health effects data is the Integrated Risk Information System (IRIS) database
maintained by the U.S. EPA. Health effects data are listed in Table 3-10.
3.3 RESULTS AND DISCUSSION
3.3.1 Baseline Risks
From our modeling of the emission and dispersion of pollutants, we estimate that a total
of about 96,000,000 persons live within a 84 km square grid system centered on at least one
sludge incinerator. In other words, a population of approximately this size lives within 40-50
km of at least one facility. Of this ]3opulation, a surprisingly large fraction (about 70 percent)
is exposed to pollutant plumes from more than one facility. Figure 3-3 shows the sizes of
populations falling within the grids of varying numbers of incinerators. For example, it shows
that about 1 % of the total U.S. population (about 2,000,000 persons) is exposed simultaneously
to pollutant plumes from 11 or more incinerators, while about 40% (96,000,000) is exposed to
at least one plume (i.e., more than zero).
We estimate that under current conditions, the annual incineration of about 0.8 million
dry metric tons of sewage sludge is responsible for an estimated incremental risk of about 0.3
to 4 cases of cancer per year. The lower of these two estimates is based on the "best estimate"
of organic emissions discussed above; the higher value is based on the "worst case" estimate of
organic emissions. Risks from exposure to metals account for approximately 0.07 cases each
year while between 0.2 and 4 cases are expected to result from emissions of organic pollutants.
Based on "best estimate" assumptions for emissions, > the highest aggregate risks from
individual chemicals are 0.09 cases/yr from aldrin, 0.08 eases/yr from dieldrin, and 0.01
cases/yr from 2,3,4,7,8-pentachlonxiibenzofuran (2,3,4,7,8-PCDF), as shown hi Tables 3-11
and 3-12. Aldrin and dieldrin were never actually detected in stack emissions, although they
were tested for. The detection limits at each test site were used as an upper bound on the
3-36
-------
Table 3-10
Health Effects Data for Pollutants from Incineration of Sewage Sludge*
Human Cancer
Potency
(mg/kg-day)-1
Arsenic
Beryllium
Cadmium
Chromium(VI)
Copper (cyanide)
Mercury
Nickel
Selenium
Zinc
1 , 2-dichlorobenzene
1 ,2-dichloroethane
trans 1,2-dichloroethene
1 ,2,3,4,6,7,8-heptacMorodibenzofuran
1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin
1 ,2,3,4,7,8-hexachlorodibenzofuran
1 ,2,3,4,7,8-hexachlorodibenzo-p-dioxin
1 ,2,3 ^J^^-heptachlorodibenzoforan
1 ,2,3,6,7,8-hexacMorodibenTO-rMlioxin
1,2,3,6,7,8-hexachlorodibenzofuran
1 ,2,3,7,8-pentachlorodibenzorunui
1 ,2,3,7,8-pentachlorodibenzo-p-diioxin
1 ,2,3 ,7,8,9-hexachlorodibenzoruran
1 ,2,3,7,8,9-hexachIorodibenzo-p-dioxin
(continued next page)
15
8.4
6.,3
42
NA
NA
1.7
NA
NA
NA
0.091
NA
1600
1600
16000
16000
1600
16000
16000
7800
78000
16000
16000
Risk Reference
Dose
(mg/kg-day)
0.0003"
0.005b
0.0005"
0.005"
0.005"
0.000086
0.02"
0.005"
0.2"
0.04
NA
0.02"
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Reference
IRIS, IRIS
IRIS, IRIS
IRIS, IRIS
IRIS, IRIS
IRIS
HEAST
IRIS, IRIS
IRIS
HEAST
HEAST
IRIS
IRIS
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
EPA
3-37
-------
Table 3-10 (continued)
Health Effects Data for Pollutants from Incineration of Sewage Sludge
Human Cancer Risk Reference
Potency Dose
toe/ke-davV1 (ma/t^y)
1 ,2,4-trichlorobenzene
2-chlorophenol
2,3,4,6,7,8-hexachlorodibenzofuEan
2,3 ,4,7, 8-pentachlorodibenzofuran
2,3,7, 8-tetrachlorodibenzofuran
2,3,7, 8-tetrachlorodibenzo-p-dioxin
2,4-dimethylphenol
2,4-dinitrotoluene
2,4-dichlorophenol
2,4-dinitrophenol
2,4,5-trichlorophenol
2,4,6-trichlorophenol
3 ,3 '-dichlorobenzidine
4-methylphenol
Acenaphthene
Acetonitrile
Acrylonitrile
Aldrin
*UXIlUi
Anthracene
Benzene
Benz(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
(continued next page)
NA
NA
16000
78000
16000
160000
NA
0.68"
NA.
NA '
NA
0.01
0.45"
NA :
NA ',
NA
0.24
17
NA
0.029
6.1
6.1
6.1
0.003
0.005"
NA
NA
NA
.NA
0.02b
NA
0.003"
0.002"
0.1"
NA
NA
0.05"
0.06"
0.01
0.00057
0.00003"
0.3"
NA
NA
NA'
NA
Reference
HEAST
IRIS
EPA
EPA
EPA
.. EPA
IRIS
IRIS
mis
IRIS
IRIS
IRIS
IRIS
HEAST
IRIS
HEAST
IRIS, IRIS
IRIS, IRIS
IRIS
mis
B(a)P
HEAST
B(a)P
3-38
-------
Table 3-10 (continued)
Health Effects Data for Pollutants from Incineration of Sewage Sludge
Benzo(k)fluoranthene
Benzole acid
Bis(2-chloiethyl)ether
Bis(2-ethylhexyl)phthalate
Bis(2-chlomisopropyl)ether
Butylbenzyl phthalate
Carbon tetrachloride
Chlordane
Chlorobenzene
Chloroform
Chiysene
Di-n-butylphthalate
Dibenz(a, h)anthracene
Dieldrin
Diethylphthalate
Ethylbenzene
Fluoranthene
Fluorene
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopentadiene
Hexachloroethane
Indeno(l ,2,3-cd)pyrene
(continued next page)
Human Cancer
Potency
(mg/kg-day)-1
6.1
NA .
1.1
0.014b
0.035
NA ;
0.052
1.3
NA
0.081
6.1
NA i
6.1
16
NA
NA
NA
NA
1.6
0.078
NA :
0.014
6.1
Risk Reference
Dose
(mg/kg-day)
NA
4b
NA
0.02b
0.04"
0.2b
0.0007"
0.00006"
0.005
0.01b
NA
O.lb
NA
0.00005"
0.8"
0.3
0.04"
0.04"
0.0008"
0.002"
0.00002
0.001"
NA
Reference
B(a)P
IRIS
IRIS
IRIS, IRIS
HEAST, IRIS
mis
IRIS, IRIS
IRIS, IRIS
HEAST
IRIS, IRIS
B(a)P
IRIS
B(a)P
IRIS, IRIS
IRIS
mis
mis
mis
IRIS, IRIS
IRIS, IRIS
HEAST
IRIS, IRIS
B(a)P
3-39
-------
Table 3-10 (continued)
«
Health Effects Data for PoUutants from Incineration of Sewage Sludge
Human Cancer Risk Reference
Potency Dose
(mg/kg-day)"1 /—-./«— j—\
Isophorone
Methyl ethyl ketone
Methylene chloride
N-nitrosodiphenylamine
N-Nitroso-di-n-propylamine
Naphthalene
Nitrobenzene
Other-pentachlorodibenzo-p-dioxin
PCBs
Pentachlorophenol
Phenol
Pyrene
Tetrachloroethylene
Toluene
Trichloroethylene
Vinyl chloride
0.0041
NA
0.0016 .
0.0049" :
7*
NA
NA
160000
7.7"
0.12"
NA
NA
0.0018
NA
0.017
0.3
\ — o* — o j J
0.2"
0.3
0.9
NA
NA
0.04"
0.0006
NA
NA
0.03b
0.6b
0.03"
0.01b
0.1
NA
NA
IRIS, IRIS
mis
IRIS, HEAST
IRIS
mis
HEAST
HEAST
EPA
mis
IRIS, IRIS
IRIS
mis
HEAST, IRIS
IRIS
HEAST
HEAST
* Human cancer potency values and risk reference doses are for inhalation exposure unless
otherwise noted.
b Oral exposure value.
B(a)P = Human cancer potency virtue for benzo(a)pyrene used for all polycyclic aromatic
hydrocarbons considered to be B2 carcinogens.
EPA = Estimating Exposure to Dio;rin-Like Compounds, U.S. EPA 19891.
HEAST = Health Effects Assessment Summary Tables, March 1992
IRIS = Integrated Risk Information System, 1992.
3-40
-------
I
•8
J!
fc
13
12
11
10
9
8
7
6
5
4
3
2
figure 3-3
Population by Number of Overlapped Plumes
Incineration of Sewage Sludge
10 10 10 10
Population Exposed to Larger Number of Plumes
8
3-41
-------
Table 3-11
Baseline Cancer Risk by Pollutant
'For Incineration of Sewage Sludge
Arsenic
Cadmium
Chromium
Nickel
1 ,2-Kiichloroethane
1 ,2,3,4,6,7,8-heptachlorodibenzofuran
1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin
1 ,2, 3 ,4,7, 8-hexachlorodibenzo-p-dioxin
1,2,3,4,7,8,9-heptachIorodibenzofuran
1 ,2,3,6,7, 8-hexachlorodibenzofuran
1,2,3,7,8-pentachlorodibenzo-p-dioxin
1 ,2,3 ,7,8-pentachlorodibenzofuran
1 ,2,3,7,8,9-hexachlorodibenzofuran
1,2,3,7,8,9-hexachlorodibenzo-p-dioxin
2, 3 ,4,6,7, 8-hexachlorodibenzofuran
2,3,4,7,8-pentachlorodibenzofuran
2,3,7,8-tetrachlorodibenzo-p-dioxhi
2,3,7, 8-tetrachlorodibenzofuran
2,4-dinitrotoluene
2,4,6-trichlorophenol
3,3'-dichlorobenzidine
Acrylonitrile
Aldrin
Benzene
Benz(a)anthracene
Benzo(a)pyrene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
HEI
Risk
7x10-*
SxlO4
lxlO-J
2x10-*
7-rlfr8
/A.IU
IxlO-7
4xlO"8
6X10-9
3xlO'7
2xlO-7
SxlO-7
SxlO-8
SxlO-8
2xlO's
4xlO"7
5x10^
IxlO"7
2xlO-g
9xlO'8
IxlO"5
IxlO4
4xlO«
9xlO'7
4x10-*
3xlO'7
2xlO'7
Mean
Individual
Risk
2X10-9
SxlO-8
IxlO-9
2xlO-10
8x10-"
5x10-"
2x10-"
C=- 1 /V10
JXIU
3xlO-12
IxlO-10
IxlO"10
2xlO-10
2x10-"
4x10-"
O— 1 A-10
JXIIT
IxlO8
2xlO-10
2X1O9
6x10-"
9xlO'12
4x10-"
SxlO9
7x1.0*'
2x10^
4xlO-10
2x10^
IxlO-10
IxlO-10
Total
Aggregate
' Risk*
2xlO3
7xlO-2
2xlO-3
SxlO4
IxlGr6
A 1 /V-5
4xlO
7xlO-s
2xl(rs
O 4 /V-i
Sxlu*
2X1O4
IxlO4
SxlO"4
2xlO-J
SxlO-5
e~-\ /v4
5x1 Vr
IxlO2
2X1O4
SxlO-3
8xlO-s
lxlO's
6xlO-5
7xlO'3
9xlO-2
2xlO-3
6x10*
SxlO-3
2x10*
2x10*
(continued on next page)
3-42
-------
Table 3-11 (continued) ;
Baseline Cancer Risk by Pollutant
For Incineration of Sewage Sludge
Bis(2-ethylhexyl)phthalate
Bis(2-chlorethyl)ether
Bis(2-chlproisopropyl)ether
Carbon tetrachloride
Chlordane
Chloroform
GhfyseHe
Dibenz(a,h)anthracene
Dieldrin
Hexachlorobenzene
Hexachlorobutadiene
Hexachloroethane
._ Indeno(l,2,3-cd)pyrene
Isophorone
Methylene chloride
N-nitrosodiphenylamine
N-nitroso-di-n-propylamine
other-pentachlorodibenzo-p-dioxin
Pentachlorophenol
PCBs
Tetrachloroethylene
Trichloroethylene
Vinyl chloride
TOTAL
HEI
Risk
3x10-"
IxlO-7
2x10-*
4X10-9
Ixltf*
3xlO'7
1 u> 1/V7
rxltr
6xlO'7
IxlO4
3xlO'7
SxlO-9
SxlO-9
7xlO'7
IxlO-10
SxlO-9
3xlO-10
8xlO'7
7x10-*
4xl(r7
exicr5
2xlO-8
4x10-'
2x10-*
6X10"4
Mean
Individual
Risk
IxlO-11
5x10-"
8xlO-13
2xlO"12
4xlO-10
IxlO"10
c in!I
5xlo
SxlO-10
6x10-"
IxlO"10
4xlO"12
2xlO-12
SxlO-19
SxlO"14
2xlO-12
1x10-"
4xlO-10
SxlO-9
2xlO-10
2XKT9
7xlO-12
2x10""
IxlO-9
2xlO-7
Total
Aggregate
Risk*
2xlO-5
7xlO-5
1x10^
2x10-*
6x10-*
2x10-*
f ins
oxlu
4x10-*
SxlO-2
2X104
SxlO6
3x10^
5xlO-»
8x10-"
3x10*
2xlO-7
SxlO4
SxlO"3
2x10-'
4xlO-3
IxlO"5
2xlO-5
ixicr3
SxlO'1
"Based on "best estimate" for emissions of organic pollutants.
3-43
-------
I
I
2
1
1
I
t*'
o o o o
ooooooood
p o o o o o
*
O tn
O VD
»-• o o o en ! ! !
_
•o
en
as
Jl<2'
It4
-------
emissions of these pollutants. The pollutant 2,3,4,7,8-PCDF, however, was detected at 3 sites.
The lifetime incremental cancer risk for the HEI is estimated to be 6xlO~* based on the "best
estimate" of emissions, or 7xlO"3 based on "worst case" estimates. Figure 3-4 shows the
estimated distribution of indMduzii cancer risk within the exposed population.
Aggregate non-cancer efforts are expressed as the number of persons exceeding the risk
reference dose (RfD) for each pollutant, or the ratio of exposure to RfD when this level is not
exceeded. As shown for organic pollutants in Table 3-13, we find that the Risk Reference Dose
for any pollutant emitted from sludge incinerators is not exceeded for any of the pollutants.
Where available, background intake of pollutant from non-sludge sources is included in the
calculations. Typical background intakes for metals are generally two or more orders of
magnitude higher than those resulting from the incineration of sludge. The most significant
contribution from-sludge is for mercury, for which the intake resulting from incineration is less
than 4% of the background level. In general, the contribution to ambient levels of these metals
is negligible, as shown in Table 3-14.
Based on methods and assumptions outlined in Chapter 2, we estimate that about 700
persons cross blood lead threshold!! each year as a result of the incineration of sludge. Exposure
to lead from the incineration of sludge also causes an estimated 100 cases of non-cancer disease
per year, of which more than 95% consist of additional cases of hypertension expected in adult
males. The remainder are due to other cardiovascular effects in adults and neurological
developmental effects in children. For cadmium, fewer than one person per year is expected
to cross the kidney cadmium threshold because of exposure from sludge.
3.3.2 Benefits from Regulatory Controls
Analysis of each facility's individual response to the regulation has been completed for
only those 23 POTWs included in the analytic subset of the NSSS. In order to estimate risk
under the new regulations, all exposure and risk calculations are repeated after adjusting
emissions from each of these modeled facilities to reflect the expected installation of additional
pollution controls. For simplicity, only two additional control technologies are considered. To
reduce emissions for metals, a wet electrostatic precipitator (ESP) operating at either high (95 %
of maximum) or medium (90% of maximum) is assumed to be installed; further reduction of
organic emissions is assumed to be unnecessary. Estimates of exposure and risk after installation
of controls are scaled to the national level based on disipersion results from the full inventory of
known incinerators, and on sample weights from the NSSS. Expected reductions in risk are
derived by subtracting these results from comparable estimates of exposure and risk under
baseline conditions. Table 3-15 shows the estimated reduction in the number of cancer cases
expected each year. Total risks from emissions of metal pollutants are expected to decline by
about 90%; risks from organic pollutants are not affected by the regulation.
Additional pollution controls reduce the number of people crossing the lead threshold
from about 700 to about 90 and the associated number of disease cases from about 100 to about
30. Table 3-16 shows how the incremental body burden of lead from incinerated sludge (in
is distributed among exposed individuals both before and after implementation of controls
3-45
-------
Figure 3-4
Baseline Individual iCancer Risk by Population Size For Incineration
10
-3
10
•£ 10
-5
10
-7
J J_
J : 1 1 L
10
10
2 3
10 10
4 5
10 10
10 10
Population Exceeding Individual Risk Level
10
8
3-46
-------
Table 3-13
•
i
Comparison of Exposure to Risk Reference Doses
For Organic Pollutants from Incineration of Sewage Sludge*
1 ,2-dichlorobenzene
1,2,4-trichlorobenzene
2-chlorophenol
2,4-dimethylphenol
2,4-dichIorophenol
2,4-dinitrophenol
Acetonitrile
Aciylonitrile
Aldrin
Benzole acid
Bis(2-ethylhexyl)phthalate
Carbon tetrachloride
Chlordane
Chlorobenzene
Chloroform
Dieldrin
Ethylbenzene
Fluoranthene
Hexachlorobenzene
Hexachlorobutadiene
Hexachlorocyclopent&diene
Hexachioroethane
Methyl ethyl ketone
Naphthalene
Nitrobenzene
Pentachlorophenoi
Pyrene
Tetrachloroethylene
Toluene
Trans 1,2-dichloroethene
Riisk
Reference
Dose
(mg/ktf-day)
0.04
0.003
0.005
0.02
0.003
O.U02
0.01
0.00357
0.00003
4
0.02
0.0007
0.00006
0.035
0.01
0.001)05
0;3
0.04
0.0008
0.002
0.00002
0.001
0.3
0.04
0.0006
0.03
0.03
0.01
0.1
0.02
Maximum
Exposure
(mg/kg-day)
ixlfr6
IxlO-5
IxlO*
2x10^
2x10*
IxlO-5
2x10-*
4xlO-s
9x10"*
2x10^
2x10"*
7x10-"
SxlO-7
2xlO-«
4x10-*
8xlO"«
4x10-*
9xlO"7
2xlO"7
IxlO-7
4xlO'7
4xlO-7
IxlO-3
2xlO-5
5x10-*
3xlO-«
SxlO-7
9xlO-«
8xlO's
IxlO'7
Maximum
Exposure
as Percent
ofRfD
0.004%
0.47%
0.02596
0.008%
0.067%
0;52%
0:19% .
. 7.5%
30%
0.005%
0.01%
0:01%
1.3%
0.038%
0.035%
16%
0.001%
0.002%
0.021%
0.005%
2;07%
0.037%
0.003%
0.041%
0.008%
0.010%
0.002%
0.09%
0.08%
0.001%
Typical
Exposure
(mg/kg-day)
7xlO-10
7xlO-»
6x10-'°
8x10-'°
7x10-'°
SxlO-9
9xlO»
2x10*
4xicr«
IxlO-7
9xlO-w
3x10-"
3x10-'°
9x10-'°
2x10^
4xlO*
2x10^
4xlQ-l°
8x10-"
5x10-"
2x10-'°
2x10-'°
5x10*
8xlO-»
2x10-"
1x10^
.2x10-'°
4x10-*
4x10^
7x10-"
Typical
Exposure
as Percent
ofRfD
<0.001%
<0.001%
<0.001%
<0.001%
<0.001%
<0.001%
<0.001%
0.003%
0.013%
<0.001%
< 0.001%
< 0.001%
<0.001%
<0.001%
<0.001%
0.008%
<0.001%
-------
II
8 K
| |>
O O O O o O g
T
-------
Table 3-15 .
\ Cancer Risks from Incineration
Before and After Regulatory Controls
For Incineration of Sewage Sludge
Baseline
With Controls Reduction
METALS*
Multiple Hearth Furnaces
Fluidized Bed Furnaces
Total
ORGANIC POLLUTANTS*-"
Multiple Hearth Furnaces
Fluidized Bed Furnaces
Total
ALL POLLUTANTS*'1"
Multiple Hearth Furnaces
Fluidized Bed Furnaces
Total
AVERAGE INDIVIDUAL RISKd
INDIVIDUAL RISK FOR HEF
0.1
0.00001
0.1
0.2-4
0.00009
0.2-4
0.3-4
0.0001
0.3-4
2xlO-7 - 3x10-*
6x10* - 7xlO-3
0.01
1x10*
0.01
0.2-4
0.00009
i 0.2-4
0.2-4
0.00009
0.2-4
2xlO'7 - 3x10-*
4x10* - 7xlO-3
0.09
9x10*
0.09
0
0
0
0.09
9x10^
0.09
2-28%e
3-39%6
•All estimates are total expected incremental risk of cancer as result of exposure to
pollutants from sludge. Estimates are in units of expected cancer cases per year.
Values do not sum to totals because of independent rounding to one significant
figure.
"Lower estimates derived with "best estimate" of emissions for organic pollutants.
Higher estimates for "worst <;ase" estimate of organic emissions.
'Estimates expressed as incremental risk that individual will develop cancer from
lifetime of exposure to pollutants from sludge.
"lUsk for average exposed in
-------
Table 3-16
»
*
Risks from Lead and Cadmium
For Incineration of Sewage Sludge
Baseline
(cases/yr)
Control
(cases/yr)
Benefit
(cases/yr)
Persons Crossing Cadmium Threshold
<0.01
<0.01
<0.01
Persons Crossing Lead Threshold
Men
Women
Children
TOTAL
Expected Disease Cases from !Lead
500
90
40
700
100
80
9
2
90
30
500
80
30
600
90
Note: all values have been independently rounded to one significant figure. Values do
not sum to totals because of independent rounding.
3-50
-------
and installation of wet ESPs. As op be seen from Figure 3-5, the average increment in blood
lead for the worst grid cell modeled would be about 0.5 ^g/dl under baseline conditions and
about 0.1 after controls.
3-51
-------
Figure 3-5
Increment to Blood Lead by Population Size for Incineration
After Controls
1 io2 io3 104 io5 io6 107 io8
Population with Larger Increment to Blood Lead
3-52
-------
4. LAND APPLICATION: DIETARY PATHWAYS
4.0 INTRODUCTION '
Each year in the United States, approximately 1.4 million dry metric tons of municipal
sludge are applied to a variety of knd types, including forests, roadsides, nurseries, pasture, and
agricultural land. Approximately one million dry metric tons of municipal sludge are applied
each year to land used for production of animal feed or human food crops. Land application,
while serving as a method for managing sludge, also serves' the beneficial purpose of fertilizing
and conditioning soil. There are several pathways, however, through which small quantities of
heavy metal and organic contaminants in the sludge might affect human health. This chapter
evaluates the application of sludge to pasture and cropland, and examines the following potential
pathways of exposure:
(1) Sludge is incorporated into the soil of farmland used for producing food crops.
Contaminants in the sludge are drawn from the soil into the tissues of those
crops, and are then ingested by humans who consume the crops directly.
(2) Sludge is incorporated into the soil of farmland used for producing animal feeds
or for pasture. Contaminants in the sludge are absorbed into the tissues of these
feeds or pasture grasses, which are then consumed by livestock. Meat and dairy
products from these livestock are consumed by humans.
(3) Sludge is applied to the surface of pasture land, and adheres to pasture grasses
Grazing cattle or lambs ingest the sludge directly as a fraction of their pasture
consumption. Humains then consume beef, dairy products, or lamb.
Section 4.1 describes the methods we use to estimate health risks for both the total
exposed population and a highly exjrased individual (HEI); Section 4.2 contains the data sources
and model inputs; and Section 4.3 presents both the results of the analysis for the baseline
(current) practices and a discussion of potential benefits from regulations. Application of sludge
to food and animal feed crops may also result in surface water, groundwater, and air
contamination. Our methods for assessing human health risks through these pathways are
developed in Chapter 5.
4.1 METHODOLOGY
4.1.1 Overview ;
To estimate health risks from land application of municipal sludge, we have created a
computer model in Borland International Inc.'s Turbo Pascal programming language, and
executed the model on an IBM-compatible personal computer. The model first calculates the
uptake of contaminants by crops and by animals feeding on crops and pasture. It then calculates
human exposure and health risks using data for human dietary consumption of these animal
4-1
-------
products and crops. Model inputs include data on sludge application rates, concentrations of
individual contaminants in the sludge, uptake rates of soil contaminants into various crop tissues
uptake rates of contaminants jn animal feed into meat or dairy products, the fraction of each type
of feed in animal diets, humah dietary data, the fraction of crops produced in sludge-amended
soil, health risk reference doses and human cancer potency slopes for sludge contaminants and
an estimate of the population exposed. Estimates of exposure and health risks for an average
individual, for the HEI, and for the entire population exposed are estimated. Separate
tabulations are provided for estimated cancer cases and for estimated risk of cadmium-related
or lead-related diseases.
4.1.2 National Versus Local Aggregation
As discussed in Chapters 3 and 7, human health risks from the incineration or surface
disposal of municipal sludge tend to be concentrated in the proximity of disposal facilities
Produce grown with land-applied sludge, however, may be distributed nationally, so the health
effects from dietary exposure to sludge contaminants are not necessarily dependent on the
location of land application practices. Although a plan€-by-plant or state-by-state analysis of land
application might consider local differences in land application practices/this analysis of health
risks has been conducted at the national level of aggregation. The discussion below outlines
some of the issues raised by the aggregation of the analysis of both food production and food
consumption to the national level.
Food Production
Municipal sludge may be applied more frequently to certain crops than to others. In
addition, land application practices may differ among states due to state regulations. However
data detailing the particular crops to which each state's sludge is applied are not available at this
time. For this analysis we therefore rely on data from the U.S. Census of Agriculture to
estimate the fraction of national food production originating in sludge-amended soils. In doing
so, we assume that the relative frequencies of crops grown with sludge do not differ appreciably
from those grown without sludge. We also assume that land application practices are uniform
across all states.
Extending the analysis of crop production to a state or regional level could provide a
more accurate estimate of the health risks associated with land application. With the availability
of data from the National Sewage Sludge Survey, such an analysis could consider plant-specific
data for quantities of sludge, application rates, and concentrations of contaminants. State-level
resolution of crop information might also point out local differences among crops grown in those
regions where land application is most common. Furthermore, since many individual states have
their own regulations or guidelines governing the land application of sludge to agricultural land,
describing statewide practices according to these restrictions would give a more accurate and
realistic picture of application patterns and crop production.
One indirect means of assessing these local differences might be the use of the
Agricultural Census, from which crop production and acreage is available at the county level.
4-2
-------
If the land application activity in any particular county reflects the relative distribution of crops
in that county, then those distributions, along with information about local sludge production
quantities and state restrictions or regulations, could be used to estimate actual farm production
by crop from sludge-amended1 soil.
Preliminary analysis of the AGDATC data base from the Oak Ridge National
Laboratory's Radiation Shielding Information Center CBaes el al, 1985) suggests that patterns
of crop production in counties where land application is ^practiced differ appreciably from
national averages. Additional investigation of these differences could offer further refinement
to estimates of dietary health risks from land application.
Food Consumption .
Data are not currently available to describe precisely how harvested food crops are mixed
and distributed throughout the United States. In the absence of information on the transport of
these crops, population risks have been computed assuming complete national mixing for all
crops grown on sludge-treated soil. This assumption neglects the possibility that certain
geographic areas might consume higher fractions of sludge-grown foods. Insofar as the
distribution network mix departs JJrom a complete national mix, non-cancer risks may be
underestimated. In particular, if food distribution is localized around an agricultural area with
a high level of land application activity, the population in that area will have a higher exposure
than estimated from this analysis, and will be more likely to exceed threshold doses for non-
cancer health effects. (Since no threshold dose is assumed for cancer, the food distribution
pattern would not affect the number of predicted cancer cases, but it would affect the
geographical distribution of those cancer cases.)
4.1.3 Description of Calculations
The risk calculations consist of four general steps. First, tissue concentrations of
contaminants in each crop as a result of the land application of sludge are modeled. Second,
concentrations of these contaminants are estimated for meat and dairy products. As discussed
above, contaminants are assumed to enter meat and dairy products either due to animal ingestion
of sludge-treated crops or direct ingestion of sludge adhering to pasture grasses! Third, the
amount of each contaminant in all crops and animal products ingested by humans, as well as the
background intakes of each contaminant are summed, to estimate total exposure. Finally, we
use the exposure estimate together with dose-response and threshold dose information to derive
an estimate of the human health risk posed by each contaminant. The application rate is then
set to zero and all calculations repeated in order to determine non-cancer risks due to
background intakes alone. Details of these calculations, are provided below.
Determining Tissue Concentrations of Contaminants in Crops Grown on Sludge-Amended
Soil
The background mass of each contaminant in the mixing zone is first calculated from the
background concentration of each contaminant and the mass of soil in the mixing zone:
4-3
-------
= MSH BS. lO'3
where:
LBj - mass of contaminant/' in mixing zone for one hectare of farmland as result
of background concentrations in soil (kg/ha),
MSH = mass of soil, in mixing zone of one hectare of farmland (Mg/ha),
BSj = background concentration (dry wt) of pollutant/' in soil (mg/kg or g/Mg),
and
10"3 = constant to convert units from (g/ha) to (kg/ha).
The mass of contaminant added through the application of sludge is calculated as:
LA,. = Nt Cj ARt ID'3
where:
mass of contaminant / added! to one hectare of farmland by land
application of sludge to crop i (kg/ha),
= total number of years sludge is applied to farmland used to grow crop /,
= concentration of pollutant/ in sludge (mg/kg or g/Mg),
= application rate of sludge for crop / (dry Mg/ha), and
10" = constant to convert units from (g/ha) to (kg/ha).
We sum these two estimates and divide by the mass of soil in the mixing zone (adjusted
for the addition of additional mass from sludge) to approximate the concentration of contaminant
in treated soil:
CT
(JV,. AR) + MSH
where:
CTjj = concentration of contaminant/ -in soil used to grow crop /, adjusted for
background :soil concentration and for additional soil mass from added
sludge (mg/kg), and
1000 = constant to convert units from (kg/Mg) to (mg/kg).
Note that as N, approaches infinity, CTV approaches C}, so repeated applications of sludge cannot
increase the estimated concentration of contaminant in soil beyond the concentration in the
sludge.
Once we have estimated the concentration of contaminant in soil, we use estimated uptake
rates to calculate the expected concentration of contaminant in crops grown on that soil:
4-4
-------
where: '
CDS = tissue concentration (dry wt) of contaminant./ in crop i (mg/kg), and
Us = rate of upfctke of pollutant j into tissue of crop i (mg/kg dry weight per
mg/kg in soil).
Determining Tissue Concentrations of Contaminants in Meat and Dairy Products Produced
from Animals Eating Feeds Grown with Sludge-Amended Soil
If farm animals are given feeds grown in sludge-amended soil, or allowed to graze on
treated pastureland, contaminants from sludge may be absorbed into animal tissues, leading to
tissues, we first derive an average concentration of each contaminant in each animal's feed mix:
- Ft CD, + FSt:Cj
where:
CFjk = weighted average concentration of pollutant j across all food sources for
animal producing meat or dairy product k (mg/kg),
Ffc = fraction of animal's food from crop i for animal product k (kg/kg or
dimensionless), and
FSfc = direct ingestion of sludge (adherence pathway) as fraction of animal's diet
for animal product k (kg/kg or dimensionless).
For cattle and lamb, this estimate includes the contribution from direct ingestion of sludge from
treated pastureland. We then use animal uptake rates to convert these results into estimated dry
weight concentrations of contaminants in each meat or dairy product:
where:
CDjfc ~ concentration of contaminant j in animal product k (mg/kg), and
Ujk = rate of uptake of contaminant y into meat or dairy product k per unit of
concentration in animal's feed (mg/kg dry* weight in animal tissue per
mg/kg dry weight feed).
Determining Individual and Population Risks from Contaminant Ingestion Through Foods
Grown in Sludge-Amended Soil
To determine human exposure to contaminants from sludge, we combine these estimated
concentrations of contaminant in food products with assumptions about dietary consumption and
the fraction of national produce grown in sludge-amended land:
4-5
-------
EXPj =
ID
'3
FCk DCk 10
'3
where:
FQ -
DC; =
FCfc =
DCfc =
10-3 =-
exposure of <;x>ntaminanty from crops, meat and dairy products produced
with sludge-amended soil (mg/kg-day),
fraction of dietary consumption of crop i grown in sludge-amended soil
(dimensionless),
daily dietary consumption of crop i' (g/kg-day),
fraction of dietary consumption of animal product k produced with sludge
(kg/kg or dimensionless),
daily dietary consumption of meat or dairy product k (g/kg-day), and
constant to convert units from (jtg/kg-day) to (mg/kg-day).
Cancer Risks
For contaminants classified as carcinogens, we convert this estimate of exposure into an
incremental risk of cancer for the exposed individual:
a.
q
where:
0, -
sum
incremental <:ancer risk from contaminant j for exposed individual
f (incremental risk of developing cancer per lifetime of exposure), and
1 j ~ human cancer potency of contaminant j (mg/kg-day)"1.
We assume that small incremental risks from individual contaminants are additive, so we
to calculate the total incremental risk for exposed individuals:
where:
CI
total individual cancer risk from dietary pathway (incremental risk of
developing caincer within lifetime as result of land application of sludge).
We multiply this result by the size of the U.S. population and divide by the human life
expectancy to estimate the incremental number of cancer cases caused annually by the
application of sludge to agricultural land:
CP =
CI POP
LS
where:
4-6
-------
CP - total aggregate risk of cancer from dietary pathway (number of additional
cancer cases expected per year as result of dietary pathways for land
application]!,,
POP = exposed population, and
LS = lifespan of average individual (yr). \
Health Risks from Non-Carcinogens
For non-carcinogenic contaminants, we express risk as the ratio of expected exposure to
the risk reference dose (RfD) for each contaminant of concern:
= POP (IF
= 0 (IF EXPj+BI.
where:
NCP, = number of people exceeding RfD of pollutanty due to sludge exposure,
BIj = background intake of pollutant; (mg/kg/day), and
RfDj = Risk Reference Dose for pollutant./ (mg/kg/day).
This calculation is repeated with iand without the application of sludge; the difference in NCPj
for these two scenarios provides a measure of aggregate risk.
The methods used for estimating non-cancer health effects from lead and cadmium differ
from those used for other non-carcinogenic contaminants, and are explained in Chapter 2.
Lead's effects are estimated separately for men, women, and children. The model bases
calculations on estimated blood lead levels, and uses nonlinear dose-response functions to
estimate non-cancer effects. Foir cadmium, expected cases of kidney disease are estimated
separately for smokers and nonsmokers. Calculations are based on kidney cadmium uptake
rates, and an individual exceeding a threshold of 200 /*g/g cadmium in the kidney is defined as
a "case". These methods do not imply that everyone exceeding this threshold will experience
kidney disease; rather, those exceeding this threshold are considered to be at risk from this
health hazard.
4.2 DATA SOURCES AND MODEL INPUTS
Assessing dietary exposures and risks from the land application of wastewater sludge
requires a number of assumptions and types of input data. These assumptions and their
ramifications are listed hi Table 4-1. This section describes the sources, assumptions, and
methods used to prepare these inpiuts for risk calculations.
4-7
-------
U Ji
4
4)
1
K
g
,o
.S
op
4
U
I
I
«!
-------
r
1
1
*
• 1
o
i I
,8. <
*< "2
*n Ui
•M (^
H CO
- §
Ig
i
to
g
I
o
I
-------
Ramification
fill
S 8.9 S
e
•8
I
I
4
o>
§
CO
U
t
60 i)
o
•§
S
-------
4.2.1 Application Rates
Wastewater sludge is.used both as a source of plant nutrients and as a soil conditioner
for agricultural applications. A]pplication rates vary according to the crop involved state
regulations, and the purpose for which the sludge is applied. For simplicity, however, the same
application rate (11 metric tons per hectare) is assumed for all crops, and sludge is assumed to
be soil-incorporated for all crops except pasture.
Although application rate would at first glance seem to be an extremely sensitive input
parameter for national population risk estimations, other assumptions in the model render the
results relatively insensitive to choice of application rate. Lowering application rates will of
course, decrease the quantity of sludge contaminants per acre available for plant uptake
However if baseline application rates are lowered while holding constant the volume of land-
apphed sludge, the amount of land to which the sludge is applied must increase proportionately
so that a larger fraction of the nation's produce is affected by land application. An increased
fraction of produce from sludge-amended soil means that a larger fraction of the average
American s diet consists of foods affected by land application.
Because of these two opposing factors, baseline estimates of average individual dose are
not significantly decreased by spreading a constant volume of sludge over an increased area of
farmland. This result relies on the assumptions that plant and animal uptakes are approximately
linear at the concentrations and exposures of interest, and that all crops and animal products
produced on sludge-amended land are mixed in the national markets.
On the other hand, if miring is not complete, decreasing the application rate and
increasing the area of application; could reduce dietary intakes of contaminants for certain
individuals who consume an above-average proportion of locally grown food. Therefore,
although the choice of baseline application rates does not affect the estimates of population risk
it does affect the estimated HEI risk, since the fraction of the HEI's diet produced from sludge-
amended soil is assumed to be independent of application rate.
4.2.2 Concentration of Contaminants in Sludge
Listed in Table 4-2 are average and 99th percentile concentrations of contaminants in
land-applied sludge, background contaminant levels in agricultural soil, and average background
contaminant intake levels from other sources. Before application of sludge, soil is assumed to
be free of the organic chemicals on the list of possible sludge contaminants, but is assumed to
contain background concentrations of many of the metals. This analysis assumes organic
contaminants in the soil decay to negligible levels between yearly harvests; therefore they do not
accumulate in soil with repeated applications of sludge. However, metals do not decay, and
might remain available for plant uptake (see, for example, Heckman et al., 1987). Although
metals may bond to the soil and become unavailable for plant uptake, or be carried away from
the topsoil by wind erosion or percolating rainwater, we conservatively assume for this analysis
that metals accumulate and remain bioavailable, and that sludge is applied for 20 consecutive
years to all crops except pasture.
4-11
-------
Table 4-2
Pollutant Concentrations in Sludge and Soil
Aldrin/Dieldrin
Arsenic
Benzo(a)pyrene
Cadmium
Chlordane
Copper
DDT/DDE/DDD
Fluoride
Heptachlor
Kexachlorobenzene
Hexachlorobutadiene
Iron
Lead
Lindane
Mercury
.Molybdenum
Nickel
PCBs
Selenium
Toxaphene
Zinc
Mean
Concentration
in Sludge
(rag/kg)1
0.021
10
11
10
0.25:
I
520
0.021
0
0.020
11
0
0
140
0.025
3.6
11
66
1.3
7.4
0:99
1,300
99th Percentile
Concentration
in Sludge
(rag/kg)'
0.10
62
67
120
1.3
2,500
0.12
0
0.10
667
0
0
490
0.13
18
51
980
6.1
49
5.1
33,000
Background
Concentration
in Soil
(rag/kg)11
0
3
0
0.2
0
19
0
0
0
0
0
0
11
0
0.1
2
18
0
0.21
0
54
Background
Intake for
Adult
(mg/day)
0
0.082
0
0.027
0
0.16
0
0
0
0
0
0
0.11
0
0.0066
0
0.17
0
0.12
0
13
• Mean and 99th pcrcentile sludge constituent concentrations obtained from the analytic survey of the
National Sewage Sludge Survey. Samples where contaminant was not detected have been assigned
values equal to the their respective limits of detection.
* Background soil concentrations from U.S. EPA (1988d).
4-12
-------
For cancer, this analysis is "concerned primarily with incremental risks from the land
application of wastewater sludge, so background soil concentrations and intakes of carcinogens
are not of particular concern'4 Background soil concentrations and intakes are important in the
estimation of non-cancer risks, however, since these risks are assumed to have thresholds (risk
reference doses or RfDs), beneath which no adverse health effect is expected. Background
concentrations and background intakes must be known to determine whether an individual's total
exposure to a contaminant will exceed the RfD when sludge is added to the soil.
4.2.3 Uptake Rates
Uptake Rates into Plant Tissue
Table 4-3 lists uptake rates for pasture and the eight food crop groups selected for this
study. Uptake rates for this analysis were obtained from plant uptake response slopes (U.S.
EPA, 1992c). The values hi Table 4-3 represent geometric means of the plant response curves
from field sludge experiments. The plant response curve for peanuts was set equal to that of
legumes. Pasture uptake values were only available for lead, copper, and molybdenum; all other
values were set equal to those for grains and cereals. All rates are expressed hi units of mg/kg
of dry weight plant tissue per mg/kg of dry soil. It is assumed that uptake rates for legumes or
grains grown for animal feed do not differ from those grown for direct human consumption.
Uptake Rates into Meat or Dairy Products
From crop uptake rates and from information on the typical diets of livestock, the
average dry weight concentrations of contaminants in animal feeds can be calculated. The rates
at which animals incorporate these feed contaminants into their tissues are then used to calculate
concentrations of the contaminants hi meat or dairy products. Table 4-4 lists uptake rates for
all meat and dairy products included in this study. The rates at which metals are taken up in
animal fat were not available, so we conservatively extend metal uptake rates from nonfat tissue
to fat tissue.
4.2.4 Animal Feed Mixes
Contaminant uptake rates differ among crops used for animal feed. The particular mix
of these crops in the diet of livestock will therefore affect the total quantity of contaminants that
animals ingest and the concentrations hi the meat or dairy products produced by these animals.
Commodity Maps from the U.S. Dspartment of Agriculture (USDA, 1982) provide information
on these feed mixes. Table 4-5 summarizes national feed totals (hi millions of metric tons per
year, mmt/year) for beef and milk cattle, boiler and layer chickens, hogs, arid lambs. From
these data, and from pasture data described below, the percentage that each type of crop
contributes to an animal's total diet can be derived.
4-13
-------
3
0)
3
CO
•s
I
I
2
.3
s
I
•u «
I
11
S
so rS
o o
oooooogo. gog
O O O O O*
®gogo!soooooogo:gogogo3
o 6 o Q < Q d o' o
— J2 !"! f _ OB
°' o' o' o o' ° d o 2
°' o' d o' o' ° o o' d
<=>' o o d :d d d d
vg
co
o g o S o g o o o o o o g o g q g o g o 3
°" d d d d o d o" d
VO 00
o o
-_- ^^r ,^ ^^, ^^,
o d d d d
-••2S 2 •« *° oe **
ogogo|oooooogog5jgogos
' ' ° ' '
c> c>
«
es o
O o
9 g
^,1
o o
O O
o o
o §
n
a
3 £ g •§ 3 ih
< -5 eS u u 0
.
- s
£u p
as&ssg|.i|fa8|i.l
QE333333Aj3SSSa(JJ^H
f
f
1
CM
1
S
•s
I
05
I
-------
i
•a
1
p
£
.a
o
3
3 * *
o •&
"9 Q
H 1
1
.1
tfl
i
t
&
£&
^
o 3
&
£*
Ife
S-31
11
n
•S ^
a
60 9
tffe
•c*
&«•
g 1.
0 i.
ll
1!
l!
&
2
~oogxo'~;o*^^'':oo2Soo'*t~it":o
§*" '» 25 <«
oooooooooooooo<"i©o
o o o
s is s * s? a s??s
5 x® m^ci cs o ci d
^ o
1 •
5 ci d d
o
»soog?o-^o~t~'»:oo'^8oofn0««ri00eS0e'*ddd
S ... oo >n oo
* SSRS!? "S a.9«3
01 o-d <* •» m ri o'd do
od^"^*c< o_; oo
0 °
oooSoooooooooogooovloo
o . • o
0 °
1 J
11 §11 e
'fiu-lll^0*!^^ «bl s"
[2'e'o''g'Eg'28'
SE i'Ss o"oj §•« « ji S.I " 2 ~ y ~s S .1
<
-------
.8
» s S a 2 8 S
in in
O\ O <-« —i
S-* ts -••*
"-1 -* -en
° ° si -
S
ON
S S g g g 3
S
S
0 «
I I 2 ""* *
-4 d d
OO
O>
i i i a a a
d 5;
21
s
o\
8 5 8
^^ cn o
o o d
I 1
S
3 a a
^j* o ^o
§ 1
U CD
i
§p
S "I * % S
I o *s 1 i
- o a. -g J -a
i i S ^ "
fc< fc X
i> &
s «
«
O O
Other
Other H
Pasture
co
b
o
in
R
•g a
I8
w'S
li.
^4
II
l>
11
||-lj
Ills
IS 8
B ,j .... .j^
5 a
4
-------
Estimates for pasture consumption (also in mmt/year) were derived from the AGDATC
date base maintained by the Oak Ridge National Laboratory. The estimated national total of
202.7 mJhon metric tons of pasture consumption per year for beef cattle, milk cattle, and lambs
was divided in proportion to'total tonnage of other feeds consumed by these animals to obtain
estimates for pasture as a fraction of these animals' diets.
Grazing animals (beef cattle, milk cattle, and lambs) also directly ingest some sludge
through its adherence to pasture grasses. We assume that the sludge has not been soS-
mcorporated, and that one third of the time, grazing animals are feeding on forage recently
sprayed with sludge (after the normal 1-3 week waiting period). The fraction of sludge ingested
by grazing animals is approximately 2.5 percent (Chaney et al., 1987) during this time
However, during the remaining two thirds of the year, the fraction of sludge ingested is assumed
to be only l percent. Therefore, on average, the grazing animals' diet is 1.5 percent sludge.
We use this value for estimating trcth national population risk and risks to the HE! In each
case, the mass of hay/pasture consumed by the livestock is reduced to accommodate the
estimated dietary percentage from sludge ingestion. Sludge ingested by animals is assumed to
have the same pollutant concentrations as the sludge had when originally land-applied- that is
none of the pollutants decay before ingestion by animals. Final livestock feed estimates are*
listed in Table 4-6. For this analysis, we assume that animalsfee
-------
Table 4-6
Estimated Feed Mix for Each Animal Product
Beef
Dairy
Eggs
Lamb
Poultry
Pork
Soil
1.5%
1.5%
0.0%
1.5%
0.0%
0.0%
Pasture
65.0%
45.2%
0.0%
74.8%
0.0%
0.0%
Soybeans
10.6%
29.7%
20.6%
5.1%
30.6%
13.0%
Grains
23.0%
23.7%
79.4%
18.6%
69.4%
87.0%
Total
100.0%
100.0%
100.0%
100.0%
100.0%
100.0%
Note: Assumes 1.5 percent of animal's diet is sludge, based on 7-21 day waiting period
between application and grazing. Values may not sum to totals because of independent
roundinp. ' *
rounding
4-18
-------
Table 4-7
Dietary Assumptions
Fraction of
Fraction of Consumption Consumption from
from Sludge-Amended Soil Sludge-Amended Soil
(Aggregate Risk)* (Risk to HEI)b
Dried Legumes
Garden Fruits
Leafy Vegetables
Non-Dried Legumes
Peanuts
Potatoes
v Root Crops
Beef (Lean)
Beef (Fat)
Beef Liver
Dairy (Non-Fat)
Dairy (Fat)
Eggs (Whole)
- Lamb (Lean)
Lamb (Fat)
Poultry (Lean)
Poultry (Fat)
Pork (Lean)
Pork (Fat)
2.1x10"*
2.1x10"*
•"^ 1 1 rt— 4
2.1XTU
2.lxlO"*
2.lxlO"*
2.lxlO"*
2.lxlO"*
2.1x10"*
2.1x10"*
2.1x10"*
2.1x10"*
2.1X10-4
2.1x10"*
2.1x10"*
2.1x10"*
2.1x10"*
2.1x10"*
2.1x10"*
2.1x10"*
2.1X10-4
2.5xlO-2
2.5xlO-2
2.5xlO'2
2.5xlO-2
2.5xlO-2
2.5xlO'2
2.5xlO'2
2.5xlO-2
9.7xlO'2
9.7xlO'2
9.7xlO-2
S.lxlO'2
S.lxlO'2
7.9xlO-2
9.7xlO'2
9.7xlO-2
l.OSxlO'1
I. OSxlO'1
- 9.7xlO-2
9.7xlO-2
Dietary
Consumption0
(g/kg-day)
0.04
0.06
1.3
0.03
0.09
0.03
0.22
0.02
0.28 ?
0.22
0.02
0.41
0.26
0.12
0.003
0.003
0.10
0.02
0.13
0.18
' Calculated from amount of sludge applied to land at an assumed rate of 11 metric tons DW/ha
b From U.S. EPA (1992c). ... . .
0 From U.S. EPA (1989h).
4-19
-------
To obtain the fraction of farmland affected by land application, we divide the estimated
annual dry tonnage of sludge for land-application to food chain agricultural land by the estimated
average sludge application rate. This calculation yields an estimate of the number of hectares
of cropland treated with sludge. This number is then divided by the total number of hectares
of non-idle farmland used to produce food and animal feed crops to estimate the percent of
farmland treated with sludge. Under the assumption that crops grown on sludge-amended soil
are no more or less productive than crops grown with other fertilizers or soil conditioners and
that these crops yield the same proportion of waste, export, or other non-food uses as would be
expected from non-sludge-amended crops, the proportion of food and animal feed farmland to
which sludge has been applied should approximately match the proportion of food production
and consumption affected by land .'application.
According to the National Sewage Sludge Survey, approximately 970,000 dry metric tons
of sewage sludge are applied annually to land for food-chain agriculture. If this sludge is
-JEPJJgdat about 11 metric tons^iectaj^an_ejtiin^
sludge each year. Dividing this number by total non-idle hectares of land used for crops and
pasture (418 million hectares) yields the estimate that approximately 0.02 percent of food-chain
agricultural land is treated with sludge each year. We use this estimate to represent the
percentage of the average American's diet that originates from sludge-amended land.
Since our estimates of FC are inversely related to our assumed application rates
increased (or reduced) sludge application rates reduce (or increase) our estimated FC values
proportionately. Population risk estimates are sensitive to the product of application rate and
FC, but relatively insensitive to the particular application rate chosen.
For the highly exposed individual (HEI), we use a different approach. This hypothetical
individual is assumed to consume am above-average fraction of his or her food from crops or
animals raised locally in sludge-amended soil. For these calculations, values of FC are taken
from Ryan (1991). Table 4-7 shows the values for both the HEI and for the national population
Because these estimates of FC for the HEI are independent of our assumed application rates'
changing the application rate would change estimated cancer risks to the HEI by equal
proportions, and would also be expected to increase risks of non-cancer health effects.
4-20
-------
4.3 RESULTS AND DISCUSSION
i»
4.3.1 Baseline Risks: Dietary Pathway
»
Table 4-8 provides estimates of cancer risks from the land application of sludge to food-
chain land under current conditions. As shown by the table, we estimate that about 4 cases of
cancer might be caused by every 10 years that sludge is land applied. Of this total risk, more
than 97 percent is caused by hexachlorobenzene and PCBs, and more than 99 percent can be
attributed with the addition of aldriia/dieldrin and toxaphene. As can be seen from Table 4-3,
organic contaminants in sludge are not assumed to be taken up into food crops in significant
quantities. It follows that concentrations of these contaminants should also be insignificant in
animal feed, so that human exposure from animal products produced with feed from sludge-
amended land should also be negligible. The remaining pathway through which this analysis
estimates human dietary exposure from land-applied sludge is for the adherence of sludge to
pasture grasses, with subsequent direct ingestion by animals. For this pathway, cancer risks for
each contaminant are determined primarily by each contaminant's concentration in sludge, its
rates of uptake into animal tissue, and its human cancer potency. As can be seen from
Table 4-4, estimated animal uptake rates for hexachlorobenzene and PCBs are relatively high.
This fact, coupled with their relatively high human cancer potencies (from Table 1-5) accounts
for their dominance of total estimated risks through dietary pathways.
We estimate that the HEI faces a risk of about 6x10^ of contracting cancer as a result
of dietary exposure to pollutants from land-applied sludge. As with total risks, this estimate is
dominated by hexachlorobenzene. Individual risks for the average individual are much lower:
about IxlO"7.
Uptake rates for animal and crop tissues are available from most metals. With the
exception of arsenic, none of the metals are considered potential human carcinogens for this
analysis. For these contaminants, non-cancer risks are of concern. As shown in Table 4-10 for
the average individual, exposure to metals from sludge is in all cases more than four orders of
magnitude lower than estimated background exposure from other sources (except for
molybdenum, for which background exposure is unknown.) For arsenic and zinc, these
background levels of exposure exceed or are approximately equal to the risk.reference dose.
Nevertheless, the contribution to exposure from sewage sludge appears to be trivial. As shown
in Table 4-10 for the HEI, the contribution to exposure from sewage sludge is in all cases less
than background exposure, and for aill metals, incremental exposure is less than five percent of
the RfD.
For lead and cadmium, we use additional methods to calculate potential health effects
from exposure to sludge. As discussed in Chapter 2, potential health effects from lead include
increased risk for hypertension, heart attack, stroke or death for men, and neurological effects
for children. As shown in Table 4-11, we estimate that about 20 persons per year might suffer
adverse health effects from dietary pathways of exposure to lead from land-applied sludge. A
higher number, about 300, are expected to cross blood lead thresholds of 7 jtg/dl for men, 10
for women, and 10 jig/dl for children. For cadmium, about 2 persons are expected to
4-21
-------
Table 4-8
Baseline Cancer Risks for Land Application: Dietary Pathway
Cancer Risk
AGGREGATE RISKS*
Aldrin/Dieldrin , QQQ4
Arsenic QQQ2
Benzo(a)pyrene Q
Chlordane Q ^
DDT/DDD/DDE 0 0002
Heptachlor 0.002
Hexachlorobenzene 03
HexacMorobutadiene 0
Lindane 9xlO-5
PCBs 0.1
Toxaphene 0 003
Total 0.4
AVERAGE INDIVIDUAL RISK" lxlo-7
INDIVIDUAL RISKS FOR HEP
•All values in incremental number of cancer cases expected per year as result of
exposure through dietary pathways from land application of sludge
Risk for average exposed individual of developing cancer from lifetime of exposure
to pollutants from sludge.
"Risk for the Highly Exposed Individual (HEI) of developing cancer from lifetime of
exposure to pollutants from sludge.
4-22
-------
3
a
.8
,«J
S jg*
•2 S
P Q
|g
|l
.IS
*
i
U
CQ
XXX
vo *•* «n
X X
CM C"l
£
i-< en
o
"x
0.
O
vo vo
O
o\
co
n
C5 O
t^l ^
p^l ^
CS CO
*-<' en
ON
X
ro
X
cs
X
c^ ° ^
-------
4
«
-------
Table 4-11
Baseline Risks from Lead and Cadmium
'Land Application: Dietary Pathways
Health Risk
CADMIUM (Persons Crossing Kidney Cadmium Threshold)
Smokers Q
0.02
Non-Smokers
Total
LEAD (Persons Crossing Blood Lead Thresholds)
Men 30Q
. Women ~Q
Children 4
Total 300
LEAD (Estimated Cases/Yr)
Men 1Q
Children g
Total _ 20
Note: Individual values do not sum to totals because of independent rounding to
one significant figure.
4-25
-------
thre8hold.cv«y 10 years M a result of dietary e^ure from land-
4.3.2 Benefits from Regulatory Controls
By limiting annual or cumulative loadings of pollutant from land-applied sludge to
agricultural land or by otherwise controlling management practices, for land application the
regulation .a likely to reduce potential dietary risks. The extent of that reduction could not be
determined from existing data and the methods used in this analysis. However, such reductions
in risk are unhkely to exceed our estimate of current (or baseline) risk. For this reason we
^ly health benefit from regulating land application of sludge to agriculuiral
^ to CMO" "" e
^ year "* about 20
4-26
-------
5. LAND APPLICATION:
GROUNDWATER, SURFACE WATER AND AIR PATHWAYS
5.0 INTRODUCTION
Chapters 4, 5 and 6 of thiis report concern the land application of sludge Chapter 4
discussed potential health risks through dietary pathways, and Chapter 6 will discuss potential
risks through dietary and direct ingestion pathways for members of households who apply sludge
to home gardens. This chapter estimates risks for the groundwater, surface water and air
pathways of exposure from land application.
5.1 METHODOLOGY
Our general strategy for analyzing these pathways is first to determine the expected
behavior of organic and inorganic contaminants loaded into soil. We begin by estimating the
fraction of contaminant likely to l>e lost through volatilization, leaching, surface runoff and
chemical degradation. These calculations, which we refer to as "mass balance," are based on
the principle that total contaminant mass is conserved; the total mass of sludge contaminant lost
to these processes or retained in the soil cannot exceed the total loading.1
After completing the mass balance calculations, we use mathematical models to predict
the movement of sludge contaminants through various environmental media. We then combine
our results with data describing the densities of human populations to estimate likely human
exposure and risk. Details of each of these steps are provided below.
5.1,1 Mass Balance
Our calculations of mass balance consider contaminant losses to surface erosion
volatilization, leaching, and degradation. As discussed in Chapter 4, the uptake of sludge
contaminants into crop tissue is of concern from the perspective of potential exposure
However, this process is of little importance to the overall loss of contaminant from soil. As
shown in Table 4-3, concentrations of contaminant in crop tissue are, at worst, of comparable
magnitude to those in sludge-amended soil, and the dry mass of crop tissue harvested from
treated land in a given year is small in comparison to the mass of treated soil (assumed to equal
approximately 2000 metric tons per hectare within the top J5 cm). For these reasons, we do
not believe plant uptake is a significant component of the mass balance for contaminant in soil.
We approximate all loss processes as first-order. In other words, we assume that rates
of contaminant loss to each process are always proportional to the concentration of contaminant
'For this analysis, we ignore the possibility that one contaminant may degrade into another
contaminant. •
5-1
-------
remaining in the soil. We begin our calculations of mass balance by estimating first-order loss
coefficients for each competing loss process.
Contaminant Losses Through Leaching
As discussed in Appendix B, a coefficient for the rate of contaminant loss to leadline is
calculated as: 5
d.
where:
~ ITr
and:
BD = bulk density of soil in mixing zone (kg/m3),
di = depth to which sludge is incorporated into soil (m),
H = Henry's Law constant for the contaminant (m3-atm/mol),
H = Henry's Law constant for the contaminant (dimensionless at specified
temperature),
KD = equilibrium partition coefficient for the contaminant (mVkg),
Kb,. = loss rate coefficient for leaching (yr1),
0. = air-filled porosity of soil, (dimensionless),
0W = water-filled porosity of soil, (dimensionless),
NR = annual recharge to groundwater beneath the land application site (m/vr)
R = ideal gas constant (8.21xlQ-5 m3-atm/mol-K), and
T = temperature (K).
Contaminant Losses to Volatilization
For organic contaminants, estimates of volatile emissions are based on equations provided
by Hwang and Falco (1986). After minor adjustments to units used in the original version:
Na = f'Q'D^ (5_1}
where:
Na - total emissions from the soil surface over time interval te. (kg/m2),
0e = effective porosity of soil (dimensionless),
Dei = intermediate variable to be defined below (mVsec),
C. = concentration of contaminant in air-fiUed pore space of treated soil
(kg/m3),
5-2
-------
intermediate variable to be defined below (m2/sec), and
duration of emissions (sec).
Hwang and Falco (1986) estirnate C. with the relation:
41H
where:
C,
41
concentration of adsorbed contaminant in treated soil (kg/kg), and
constant to convert units from (atm-mVrnol) to (m3/m3) at approximately
298 K.
Of interest for these calculations is the relationship between the total concentration of
contaminant in treated soil (in dissolved, adsorbed or vapor phase) and the concentration in
vapor phase within the soil's pore space. As discussed in Appendix A, we use the more
appropriate equation:
Ca = CJ [BDKD/H"+ SJH + 6J
where:
Ct = total concentration of contaminant in treated soil (kg/m3).
Other variables used within Equation 5-1 are:
and:
where:
and:
D
ei
' 1 + KS
K = PssKDfH
and where:
1Q-4
Da
P,,
constant to convert units from (cm2) to (m2),
the molecular diffusivity of contaminant in air (cm2/sec), and
particle density of sludge-soil mixture (kg/m3).
5-3
-------
These equations provide an- estimate of total emissions from an uncovered layer of
contaminated soil as a function of time and the initial concentration of contaminant. As is
evident from Equation 5-1, hgwever, the estimated loss rate is riot proportional to contaminant
concentration. For consistency with methods used to estimate losses for other pathways,
Equation 5-1 is evaluated for te equal to 1 year &=3.2xl07 sec), and results are used to
estimate an approximate loss coefficient. Losses predicted for the first year (Nay) are divided
by the total mass of contaminant in soil to estimate the approximate fraction of available
contaminant lost per unit of time. For a unit concentration! (1 kg/m3) of the contaminant in soil,
the mass of contaminant beneath one square meter of soil surface (in kg/m2) is equal to the
volume of treated soil beneath a square meter of surface (m3 per m2), which is equal to the depth
of incorporation (m). The estimated loss rate (in kg/m2-yr) is converted to a comparable first-
order loss coefficient (in yr1) as:
Na
K^, « -In(l— -*) (5-2)
di
where: ;
KVOI = loss rate coefficient for volatilization, used to approximate loss function
described by Equation 5-1 (yr1),
Na,, = emissions from the soil surface in first year (kg/m2-yr), and
dj = depth of incorporation for sludge (m, equivalent to kg/m2 for a unit
concentration of contaminant in treated soil).
Because Equation 5-1 was derived by assuming the column of contaminated soil is of infinite
depth, it can predict greater than 100 percent loss within a year for a relatively shallow layer
of treated soil and a relatively volatile contaminant. For such cases, Equation 5-2 cannot be
evaluated and the rate coefficient: is estimated from predicted emissions in the first second
('.=!):
where:
Na, = emissions from the soil surface in first second (kg/m2-sec), and
3.2xl07 = constant to convert (sec'1) to (yr1).
Contaminant Losses to Erosion
Annual losses to erosion are calculated based on an average rate of soil loss for
agricultural land. If contaminant is evenly incorporated into the zone of incorporation, a
coefficient for losses to erosion can be calculated as:
An,
where:
5-4
-------
- rate coefficient for loss of contaminant to erosion from treated land (yr1),
and
de = depth of soil eroded from site each year (m/yr).
t
Individual Loss Processes as Fraction of Total Loss
These three loss rate coefficients are combined with an estimated loss coefficient for
degradation of the contaminant in treated soil (obtained for each contaminant from scientific
literature) to yield a coefficient for the total rate at which the contaminant is lost from soil:
where:
= degradation rate coefficient for land application site (yrl), and
= total loss rate for the contaminant in treated soil (yrl).
The ratio of each individual coefficient to the total then describes the fraction of
contaminant loss caused by each individual process:
flee ~ ^lee / ^M fvol =
Jem ~ ^ero ' ^tot fdeg ~ ^deg I
where:
fte = fraction of total loss caused by leaching (dimensionless),
fvoi = fraction of total loss caused by volatilization (dimensionless),
fero = fraction of total loss caused by erosion (dimensionless), and
fdtg = fraction of total loss caused by degradation (dimensionless).
5.1.2 Estimating the Concentration of Contaminants in; Groundwater
After completing the mass balance calculations describe above, we use two additional
steps to calculate the concentration of each contaminant in groundwater near the site:
1) Determine the concentration of contaminant in water leaching through the treated
soil, and
2) Use mathematical models for the transport of contaminant through the unsaturated
and saturated soil zones to estimate expected concentrations of contaminant in
groundwater.
Calculations for the first of these two steps differ according to whether the contaminant
of concern is organic or inorganic. For organic contaminants, we conservatively assume that
sludge is applied to the land indefinitely. Concentrations of organic contaminants gradually
5-5
-------
increase in the soil until the rates of annual loss equal rates of annual loading, and steady state
annua oa
is achieved. Our calculations of aggregate risks are based on this steady
WhiCh ^ «"*»*»* laches from the site can be determined
F4; = 0.001
where:
FA, = annual flux of contaminant leaching from the treated land (kg/ha-vr)
U.OU1 = constant to convert units from (g/ha-yr) to (kg/ha-yr),
AR - application rate for sludge (dry Mg/ha-yr),
C = concentration of contaminant in sludge (mg/kg), and
fie. = fraction of, total contaminant losses attributable to leadline
(dimensionless). fe
in»* ta aSSUMed t0 to ^^ for 20 «»«»*« years, followed by an
inactive period in which contaminant is depleted from the.treated soil by leaching and erosion
To sunulate potential contamination of groundwater for metals, the loading of cxjntamin^t into
Ae unsaturated zone is "linearized" into a pulse of constant magnitude to 4«<££S£
SS^TTT?^^^0"^
The duration of that pulse is calculated so that contaminant mass is conserved For land
application sites, the maximum rate of loss is expected in the year immediately foUowine the last
application of sludge, since the concentration of contaminantat the site reach* 'tepXtS
* c> *" ^ loss rate ^ to maintained for
TP=NI{\-e{~
where:
TP = duration of "square" wave for approximating the loading of contaminant
into the top of the .unsaturated soil zone (yr), and
N = number of consecutive years sludge is applied to site (yr).
1- hin1?8 ieSUlt ^
-------
For both organic and inorganic contaminants, this flux can be combined with the assumed
rate of net recharge to groundwater at the land application site to derive an estimate of the
average concentration of contaminant in this leachate:
c _
""'
n
where:
Cfc,. = concentration of contaminant in water leaching from the treated land
(mg/l),
0.1 = constant to convert units from (kg/ha-m) to (mg/l), and
MR = net recharge in treated area (m/yr).
The next step is to relate this concentration to the expected concentration of contaminant
in drinking water wells near the site. Two mathematical models are combined to calculate an
expected ratio between these two concentrations. The Vadose Zone Flow and Transport finite
element module (VADOFT) from the RUSTIC model (U.S. EPA, 1989b,c) is used to estimate
flow and transport through the unsaturated zone, and the AT123D analytical model (Yeh, 1981)
is used to estimate contaminant transport through the saturated zone.
VADOFT allows consideration of multiple soil layers, each with homogeneous soil
characteristics. Within the unsaturated zone, the attenuation of organic contaminants is predicted
based on longitudinal dispersion, act estimated retardation coefficient derived from an equilibrium
partition coefficient, and a first-order rate of contaminant degradation. The input requirements
for the unsaturated zone module include various site-specific and geologic parameters and the
rate of groundwater recharge in the area of the site. We assume that the flux of contaminant
mass into the top of the unsaturated zone beneath a land application site can be represented by
results from the mass-balance calculations described above. Results from analysis of the
unsaturated zone give the flow velocity and concentration profiles for each contaminant of
interest. These velocities and concentrations are evaluated at the water table, converted to a
mass flux, and used as input to the; AT123D saturated zone? module.
......... The flow system in the vertical column is solved with VADOFT, which is based on an
overlapping representation of the unsaturated and saturated zones. The water flux into the
unsaturated zone is specified for the bottom of the zone of incorporation for sludge. In addition,
a constant pressure-head boundary condition is specified for the bottom of the unsaturated zone
beneath the land application site. This pressure-head is chosen to be consistent with the expected
pressure head at the bottom of the saturated zone. Transport in the unsaturated zone is
determined using the Darcy velocity (Vj) and saturation profiles from the flow simulation. From
these, the transport velocity profile! can be determined.
Although limited to one-dimensional flow and transport, the use of a rigorous finite-
element model in the unsaturated zone allows consideration of depth-variant physical and
chemical processes that would influence the mass flux entering the saturated zone. Among the
more important of these processes are advection (which is a function of the Darcy velocity,
5-7
-------
saturation and porosity), mass dispersion, adsorption of the leachate onto the solid phase and
both chemical and biological degradation.
To represent the variably saturated soil column beneath the application site the model
discretizes the column into a finite-element grid consisting of a series of one-dimensional
elements connected at nodal points. Elements can be assigned different properties for the
simulation of flow in a heterogenous system. The model generates the grid from user-defined
zones; the user defines the homogeneous properties of each zone, the zone thickness and the
number of elements per zone, and the code automatically divides each zone into a series of
elements of equal length. The governing equation is approximated using the Galeridn finite
element method and then solved iteratively for the dependent variable (pressure-head) subject
to the chosen initial and boundary conditions. Solution of the series of nonlinear simultaneous
equations generated by the Galerkin scheme is accomplished by either Picard iteration a
Newton-Raphson algorithm or a modified Newton-Raphson algorithm. Once the finite-element
calculation converges, the model yields estimated values for all the variables at each of the
discrete nodal points. A detailed description of the solution scheme is found in U.S. EPA
(1989b).
One-dimensional advective-dispersive transport is estimated with VADOFT based on the
estimated mass flux of contaminaat into the top of the soil column, and a zero concentration
boundary condition at the bottom of the saturated zone. The resulting mass flux from the
VADOFT simulation is used as input for the AT123D model, which simulates contaminant
transport through the saturated zone. It is represented as a mass flux boundary condition applied
over a rectangular area representative of the land application site. The transient nature of the
flux into the saturated zone is represented by time-dependent levels interpolated from the results
generated by the VADOFT simulation.
As in calculations for the unsaturated zone, degradation of organic contaminants is
assumed to be first-order during transport through the aquifer. Speciation and complexation
reactions are ignored for metals, leading to the possible over- or underestimation of expected
concentrations of metals in groundwater at the location of a receptor well. Detailed descriptions
of the AT123D model are provided by U.S. EPA (1986h) and by Yeh (1981) and will not be
repeated here. In general, the model provides an analytical solution to the basic advective-
dispersive transport equation. One; advantage of AT123D is its flexibility: the model allows the
user up to 450 options and is capable of simulating a wide variety of configurations of source
release and boundary conditions. For the current application, AT123D uses the source term
from VADOFT and other input parameters to predict concentrations of contaminant within 300
years in a receptor well at the downgradient edge of (the site.
5.1.3 Estimating the Concentration of Contaminants in Ambient Air
Two steps provide an estimate for the concentration of volatilized contaminants in air
near the site:
5-8
-------
1) Use the mass balance calculations summarized above to determine the mass of
contaminant expected to volatilize from the land application site within a period
equivalent to a human lifespan, and
i
2) Use a simplified version of the Indusitrial Source Complex Long Term Model
(ISCLT) to model ihe transport and dispersion of contaminant in ambient air near
the site.
With methods analogous to those for organic contaminants in groundwater, we first
estimate the rate at which contaminant mass will volatilize from the site, based on the
assumption that equilibrium has fcieen achieved between annual contaminant loadings and total
losses:
FA = 0.001
where:
FAy = annual average flux of contaminant volatilizing from the treated land
(kg/ha-yr), and
0-001 = constant to convert units from (g/ha-yr) to (kg/ha-yr).
The next step is to relate releases of volatilized contaminant to the expected
concentrations in ambient air. The model we use to simulate transport of contaminant from
treated land is described by U.S. EPA (1986h) and is based on equations provided by
Environmental Science and Engineering (1985). These equations are simplifications of equations
used in ISCLT. The exposed individual is assumed to live at the downwind edge of the land
application site. A source-receptor ratio is calculated to relate the concentration of contaminant
in ambient air at that individual's location (g/m3) to the rate at which that contaminant is emitted
from the treated soil (g/m2-sec):
9SRR = 2.032
(r1 + xp u az
where:
SRR = source-receptor ratio (sec/m),
2.032 = empirical constant,
A = area of land application site (m2) ,
v = vertical term for dispersion of contaminant in air (dimensionless),
r/ — distance from center of the land application site to the receptor (m),
Xy = lateral virtual distance to receptor location (m),
u = wind speed (m/sec), and
-------
source to the land application site, such that the angle 0 subtended by the site's width is 22 5°
This distance is calculated as: '
The standard deviation of the vertical distribution of concentration (a,) is defined by an
atmospheric stability class and the distance from the center of the site to the receptor. Table 5-1
provides values for two parameters!, a and b, for a range of distances under stable atmospheric
conditions. Based on values from this table, an appropriate value of
-------
Table 5-1
Parameters-Used to Calculate 60.00
a
15.209
14.457
13.953
13.953
14.823
16.187
17.836
22.651
27.084
34.219
b
0.81558
0.78407
0.68465
0.63227
0.54503
0.46490
0.41507
0.32681
0.27436
0.21716
1 Source: Environmental Science and Engineering (1985).
az calculated as at =evt where x is distance in km.
5-11
-------
100° AR
where:
1000
AR
average concentration of contaminant on eroded soil from the sludge
management area (mg/kg),
constant to convert units from (g/kg) to (mg/kg),
application rate for sludge to treated land (dry Mg/ha-yr), and
mass of soil eroding annually from one hectare of sludge management area
(kg/ha-yr).
* H ^ er°Si0n fiom ^ Site ^ *» calculated from the bulk density of treated soil
and the estimated average rate of soil loss for agricultural soil in the U.S:
ME
- 10 , 000 de BD
where 10,000 is a constant to convert units from (kg/m2) to (kg/ha).
, f°r metals- * dhcussed m APP^ C' concentrations
- H "7 ass"medAto mcrease from y<** to year as contaminant accumulates in
the soil and then to decrease after the last application of sludge. Since human exposure
continues for the duration of an individual's lifetime, expected concentrations of metal
contamman s in surface water are calculated based on maximum estimated average losses of
contaminant through surface erosioim for a period equal to the human life expectancy.
WhCh SludgC ^ applied'
is assumed to be loaded once per
vM , - ' n
year at an arbltrary rate of 1 kg/ha-yr, and lost at a continuous first-order rate (fj.
outcome of combined loading and losses is calculated numerically as:
where:
N
mass of contaminant in soil at end of year t (kg/ha),
lumped rate coefficient for contaminant loss from treated land (yr1), and
number of years in which sludge is applied (yr).
°f contaminant to
Pkce> but contaminant
where:
MN =
MLS =
mass of contaminant in soil at end of a period equal to an individual
lifetime (kg/ha),
mass of contaminant in soil after N applications of sludge (kg/ha), and
average human lifespan (yr).
5-12
-------
The fraction of total, cumulative toading lost in the human lifespan is independent of the
assumed application rate, and can be calculated as:
•
>
f _ 1 LS
J IS ~ A TT~-
# 1
where:
fLs - fraction of total cumulative loading lost in human lifetime (dimensionless)
1 = arbitrary unif: loading rate (kg/ha-yr). '
Our estimation of average contaminant concentrations on eroded soil for a metal is
identical to the corresponding step for organic contaminants, except that the mass of soil eroding
per year is multiplied by the life ex]3ectancy to calculate the total mass of soil lost in that period
We multiply our estimated total loading of contaminant byi the fraction expected to be lost to
erosion, and divide by the total mass of eroded soil to calculate the expected average
concentration of each inorganic contaminant on eroded soil:
1000
-i US
•»r>*4
where:
C^, = the concentration of contaminant in soil eroding from the sludge
management urea (mg/kg), and
1000 = constant to convert units from (g/kg) to (mg/kg).
The next step is to determine the extent to which eroded soil from the sludge management
area is diluted by soil from the (untreated) remainder of the watershed. A "dilution factor"
describes the fraction of the stream's sediment originating in the land application site:
where:
A^ = area of land treated with sludge (ha),
Aw, = area of the watershed (ha),
Df = dilution factor (dimensionless),
S»» = sediment delivery ratio for the land treated with sludge (dimensionless)
Sw. = sedunent delivery ratio for the watershed (dimensionless),
ME^ = estimated rate of soil loss for land treated with sludge (kg/ha-yr), and
ME*, - estimated rate of soil loss for the watershed, (kg/ha-yr).
The sediment delivery ratio for the sludge management area (SMA) is calculated with the
following empirical relationship for delivery of eroded soil from the site to the stream (U.S.
EPA, 1986h):
Ssaa = °- 77 [L^] -0-22
where:
5-13
-------
buffer zone or distance between the SMA and the receiving water body
(m).
The sediment delivery ratio for the watershed is calculated as (Vanoni, 1975):
$„ =0.872 UWJ -0-125
If we assume the rates of soil erosion from the SMA and remainder of the watershed are the
same, ME^ and ME^ cancel from the equation, and the dilution factor can be calculated as:
A
£) _ ___ _ ana
f
If all contaminant in stream sediment is assumed to originate in the land application site this
same fraction also describes the ratio between the average concentration of contaminant in
sediment entering the stream and the average concentration in soil eroding from the site:
- DF C^
where:
C«d = dry weight concentration of contaminant in eroded soil (me/kg)
.. «, _„, . _ _„ ..,.„,, _,. , „ -..^.. , . \ o 7c*X*
We use this estimated concentration of contaminant in the stream's sediment to calculate
the expected concentration of contaminant in the water column. Once the eroded soil enters the
stream, contaminant is assumed to partition at equilibrium between the solid and liquid
compartments of the stream. The dissolved concentration is related to the concentration of
contaminant on the eroded soil entering the stream as:
where:
Cw = the concentration of contaminant in surface water (mg/0,
KDW = partition coefficient for contaminant in the stream (//kg),
PI = fraction liquid hi the water column, (kg/kg),
P, = fraction solids in the water column (kg/kg), and
Pw = density of pure water (assumed to equal 1 kg//).
For metals, a partition coefficient for the contaminant is derived with an equation from U S
EPA(1982b): ' '
= ttTSS*
where:
5-14
-------
TSS = total suspended solids content of the stream (mg/Q, and
a, @ = contaminant -specific empirical constants.
The ratio P/PS is calculated as:
pi
Ps TSS
where:
10"6 = constant to convert units from (nig) to (kg).
5.1.5 Estimating Human Exposure and Risks
To estimate human exposure we use methods discussed in Sections 5.1.1 through 5.1.4
to estimate concentrations of each contaminant in groundwater, surface water and air near each
land application site. Our calculations are based on average measured concentrations of each
contaminant in sewage sludge of land-applying POTWs in the analytic survey of the National
Sewage Sludge Survey (NSSS), combined with our assumptions about local geological
hydrological, and meteorological conditions. '
The NSSS provides data for the quality and quantity of sludge applied to land, but lacks
information describing characteristics of the individual sites at which sludge is applied The
survey also does not provide information about the number of sites treated each year by
individual POTWs. For simplicity and for lack of site-specific data, we have chosen a single
reasonable worst case" prototypiciri site to represent sludge management areas for each of the
thousands of POTWs believed to practice land application. Based on characteristics of this
prototypical site (to be described below), we use mathematical models to predict concentrations
of pollutants in groundwater, surface water and air near the site.
these estimated concentrations in environmental media are converted to estimates of
human exposure based on assumptions about the rate at which the average individual consumes
drinking water and inhales air. For air, we calculate human exposure as:
EXP, = 10 C
'
,
J BW
where: "
EXPj = human exposure to pollutant j (mg/kg-day),
10'3 = constant to convert units from (fig) to (mg),
C»ir = concentration of contaminant j in air 0*g/m3),
I. = inhalation volume (m3/day), and
BW = average body weight (kg).
We calculate potential exposure through ingestion of contaminated groundwater and
surface water as
5-15
-------
and:
BW
where:
!„ = quantity of water ingested daily (//day),
Cvd = concentration of contaminant in well water (mg//), and
G.W = concentration of contaminant in surface water (mg//).
We calculate exposure through ingestion of fish based on our estimates for the
bioaccumulation of contaminant in fish and the assumed rate of fish ingestion. Bioaccumulation
is the process by which aquatic organisms accumulate contaminants, from both water and food
at concentrations higher than the ambient concentration. The process by which a compound is
absorbed from water through gill membranes or other external body surfaces iVcalled
bioconcentration, and the measure of a chemical's tendency to bioconcentrate is described by
the bioconcentration factor. Biomagnification, in contrast, denotes the process by which the
concentration of a compound increases in different organisms occupying successive trophic
levels. The combined accumulation from these two sources is represented by the
bioaccumulation factor, which is calculated as the product of the bioconcentration factor and a
food chain multiplier. This product: is multiplied by the concentration of contaminant in surface
water and the ratio of concentration in fillet to that of the whole fish to calculate exposure-
«. . - ".,-, (5-3)
where:
BCF = contaminant-specific bioconcentration;factor (//kg),
FM = contaminant-specific food chain multiplier (dimensionless),
If = daily consumption of fish (kg/day), and
Pf = ratio of contaminant concentration in fillet to whole fish (dimensionless).
To complete our calculations of risk we combine estimates of individual exposure through
groundwater, surface water, ambient air and fish with estimates for the sizes of exposed
populations. Based on an estimated density (persons per heptare) for human populations living
near land application sites and an estimate for the area of land affected by each exposure
pathway, we derive an estimate for the number of persons potentially exposed at each site This
value is combined with estimates of individual exposure and risk at each site to derive an
estimate of aggregate risk. Finally, we scale by the estimated number of land application sites
in the U.S. to calculate our estimates of total risk.
5-16
-------
5.2 DATA SOURCES AND MODEL INPUTS
*
Within the analytic survey of the NSSS, about 100 POTWs report applying sludge to
agricultural land, forests, public land, reclaimed land, dedicated land, and undefined land uses.
Based on sample weights from the survey, these sites represent an estimated 4803 POTWs. As
mentioned above, we use a single idealized prototype (based on reasonable worst-case soil,
hydrogeologic, meteorologic, and other site conditions) to represent all land application sites.
We first estimate risks for this representative site, and then scale results to the national level.
Because our analysis is based on conservative assumptions for many key parameters, actual risks
to the exposed population are unlikely to exceed those estimated here.
This section discusses each of the key input parameters we use to model land application
sites. These include: concentrations of contaminants in sludge, site parameters, soil parameters,
hydrologic parameters, chemical-specific parameters, and parameters describing human
-populations;
5.2.1 Sludge Parameters and Site Parameters
To estimate exposure and risk we begin by characterizing the concentrations of
contaminants in land-applied sludge, and the sites to which the sludge is applied.
Concentrations of Pollutants in Sludge
Table 5-2 lists mean and 99th percentile concentrations of contaminants for POTWs
reporting land application of sludge in the analytic survey of the NSSS. We use the mean values
for our calculations of total (aggregate) risk, and for our calculations of individual risk for the
average exposed individual. We mse the 99th percentile concentrations for calculating risks for
the Highly Exposed Individual (HEI).
Area of Land Application Site
The areas of individual land application sites used by plants in the NSSS are not known.
We make the simplifying (probably conservative) assumption that each POTW applies all of its
land-applied sludge to a single contiguous area in a given year, and calculate an expected size
for that area based on plants in the NSSS. We calculate this value from the estimated total
quantity of sludge used for land application (1,499,861 dry; metric tons per year), the estimated
number of treatment works using land application (4803), and an average application rate for
agriculturally applied sludge (11 mt/ha). From these values we calculate that the average POTW
treats about 28 hectares of land per year, and use this estimate as the size of our prototypical
site. This result, together with other assumed characteristics of the land application site, is listed
in Table 5-3.
Area of Watershed
We assume the area of the watershed is 440,300 ha, the mean size hydrologic cataloguing
unit as defined by the U.S. Geological Survey, (U.S. EPA, 1990a). We also assume the entire
5-17
-------
Table 5-2
Pollutant Concentrations in Land-Applied Sludge
•
Aldrin/Dieldrin
Arsenic
Benzo(a)pyrene
Chlordane
Copper
DDT/DDE/DDD
Fluoride
Heptachlor
Hexachlorobenzene
Hexachlorobutadiene
Iron
Lead
Lindane
Mercury
Molybdenum
Nickel
PCBs
Selenium
Zinc
Source: Mean and 99th percentile
Mean Concentration
in Sludge
(mg/kg)
0.021
10
11
0.25
520
0.021
0
0.020
11
0
0
140
0.025
3.6
11
66
1.3
'"_;.- 7.4,'"".. " .••-_.
1,300
sludge constituent concentrations
99th Percentile
Concentration in
Sludge (mg/kg)
0.10
62
67
4_jJtt.f_\_
120
1.3
2,500
0.12
0
0.10
667
0
0
490
0.13
18
51
980
6.1
49 1
33,000
obtained from the
analytic survey of the National Sewage Sludge Suivey.
5-18
-------
Table 5-3
*
ft
Site and Sludge Parameters for Land Application
Sludge Management Area (ha) 28*
Watershed Area (ha) 440 3QQb
Depth of Incorporation for Sludge (in) 0.15e
Lateral Distance to Well (m) 0 1000
Width of Buffer Zone (m) 10
Wind Velocity (m/sec) 4.5
-------
watershed is used for agricultural purposes, so the topography, cover, and other characteristics
of the remainder of the watershed are identical to those of the sludge management area.
Depth of Incorporation *
We assume sludge is incorporated into treated soil to a depth of 6 inches or 15 cm (U.S
EPA, 1986h).
Distance to Well
As a worst-case exposure scenario, we estimate the concentration of contaminants in well-
water at the downgradient edge of the site. We then conservatively extend exposure to
concentrations at this location to all individuals residing within a 90 degree angle centered on
the downgradient direction of groundwater flow and within a distance of 1 km. Similarly we
extend estimated concentrations at 1 km downgradient from the edge of the site to all individuals
residing within the same 90 degree! angle and at distances ranging from 1-3 km.
Distance to Human Receptor (Inhalation Pathway)
As a worst-case exposure scenario, we assume that all persons residing within 1 km of
the site are exposed to ambient air concentrations estimated at the down-wind edge of the land-
application site, and that persons residing between 1 km and 3 km from the site are exposed to
ambient air concentrations estimated at 1 km downwind.
Width of Buffer Zone
We assume that the width of the "buffer zone" between the land application site and the
nearest stream is 10 meters (U.S. ,]EPA, 1986h). A buffer zone of identical width is assumed
to separate surface water from the remainder of the watershed.
Velocity of Wind at Ground Surface
Wind velocity affects the transport of volatilized contaminant from the site We have
selected a value of 4.5 m/sec (10 raph) for wind velocity at the ground surface. This value is
based on mean annual wind speeds in the United States (U.S. EPA, 1990c).
Air Temperature
The model air temperature of 15°C represents the annual average for the U S (US
EPA, 1986f).
Application Rate
A typical application rate for agricultural utilization is 11 mt/ha (U.S. EPA, 1983b). We
use this value for our calculations.
5-20
-------
Number of Applications of Sludge
For organic contaminants, calculations are based on annual loadings of contaminant to
treated land, and sludge is assumed! to be applied indefinitely (year after year). For metals, we
assume sludge is applied to each site for 20 years. This value is consistent with the "useful life
of application sites" as described by U.S. EPA (1983b).
Yearly Loss of Soil to Erosion
The model erosion rate of 0.00060 m/yr represents a weighted average of annual soil loss
rates presented in the U.S.D.A. Soil Conservation Survey National Resources Inventory
Summary Report (USDA, 1987). This average value lias been converted from 3.8 tons/acre-yr
based on an assumed bulk density of 1400 kg/m3 for treated soil.
5.2.2 Soil Parameters
The unsaturated zone is characterized by pore space containing both air and water,
whereas pore space in the saturated zone contains water only. Because of differences in flow
regimes, these two zones require different equations and input parameters for tracking
contaminant transport. A simplifying assumption used for these calculations is that the basic soil
characteristics (including soil type, porosity, and bulk density) of the two zones are identical.
Values for parameters describing the mixing, unsaturated, and saturated soil zones are provided
in Table 5-4.
Soil Type
The types of soil in the mixing zone, unsaturated, and saturated zones affect the ability
of a contaminant to move vertically to the aquifer and laterally to a nearby well. In general, the
ease of contaminant transport through a soil (ignoring the adsorption properties of the soil) is
largely affected by the type of clay present, the shrink/swell potential of that clay, and the grain
size of the soil. The less the clay shrinks and swells and the smaller the grain size of the soil,
the more difficult it is for contamimnts to move through that soil. Soil types in the unsaturated
zone in order of increasing ease of transport are (1) non-sririnking clay, (2) clay loam, (3) silty
loam, (4) loam, (5) sandy loam, (6) shrinking clay, (7) sand, (8) gravel, and (9) thin or absent
soil (U.S. EPA, 1985d).
Sand has been selected as a reasonable worst case soil to use in model scenarios for these
calculations. Wherever possible all values for parameters describing soil characteristics for
model simulations are based on values estimated for sand.
Porosity of Sludge/Soil
Porosity is the ratio of the void volume of a given soil or rock mass to the total volume
of that mass. If the total volume is represented by Vt and the volume of the voids by Vv, the
porosity can be defined as fl^v/V,. Porosity is usually reported as a decimal fraction or
5-21
-------
Table 5-4
Sloil and Hydrologic Parameters
For Land Application
1
SoU Type . Sand
Porosity of Sludge/Soil 0.4
Bulk Density of Sludge/Soil (kg/m3) 1400b
Bulk Density of Soil in Unsaturated and Saturated Zones (kg/m3) 1600°
Saturated Hydraulic Conductivity of Soil (m/hr) 0.61d
Water Retention Parameters
V* 0.045e
a (m-1) 14 5.
0 2.68e
Fraction of Organic Carbon in Soil or Sludge
Mixing Layer 0.01
Unsaturated Zone 0.001f
Saturated Zone : 0.0001
Depth to Groundwater (m) 1«
Net Recharge or Seepage (m/yr) 0.5h
Thickness of Aquifer (m) is
Hydraulic Gradient 0.005'
Total Suspended Solids in Surface Water
* Todd (1980), Carsel and Paimsh (1988).
b Chaney (1992).
c Based on porosity of 0.4. Freeze and Cherry (1979).
d 95th percentile value from Carsel and Parrish (1988).
e Mean values reported for sand in Carsel and Parrish (1988).
f ! ower bound of range reported in U.S. EPA (1986f).
e Conservative value.
h Average of range reported in U.S. EPA (1986f).
I Average value for groundwater surveyed in U.S. EPA (1986f).
1 Geometric. mean of values reported in U.S. EPA's STORET data base, U S EPA
(1992a).
5-22
-------
percentage, and ranges from 0 (no pore space) to 1 (no solids).
For this analysis, we assume a total porosity of 0.4, taken from Todd (1980). This value
is consistent with the average value for sand (0.43) reported in Carsel and Parrish (1988). It
is used to represent total porosity within the soil treated with sludge, and within the unsaturated
and saturated soil zones beneath the land application site.
Effective porosity is calculated as the difference between the average saturated water
content and the approximate average residual water content, and refers to the amount of
interconnected pore space available for fluid flow. For these calculations, the average residual
water content in the unsaturated zone is assumed to be less than 0.05, and effective porosity has
been approximated with the same value used for total porosity (0.4) in mass balance and
groundwater transport calculations.
Bulk Density of Sludge/Soil
The bulk density of soil is defined as the mass of dry soil divided by its total (or bulk)
volume. Bulk density directly influences the retardation of solutes and is related to soil
structure. In general, as soils become more compact, their bulk density increases. Bulk density
can be related to the particle density and porosity of a given soil as:
BD - pw(l - 6,)
where:
BD = bulk density of soil (kg/m3),
PK = particle density of soil (kg/m3), and
Qt — porosity of soil (dimensionless).
Typical mineral soils have particle densities of about 2650 kg/m3 (Freeze and Cherry,
1979). This value and a soil porosity of 0.4 suggest a bulk density of about 1600 kg/m3 for pure
soil, somewhat higher than the 1300-1500 kg/m3 range typically encountered for soil mixed with
sludge (Chaney, 1992). We assume the bulk density of the soil in the mixing zone is 1400
kg/m3, and the bulk density of soil in the unsaturated and saturated zones is 1600 kg/m3.
Saturated Hydraulic Conductivity of Soil
Saturated hydraulic conductivity refers to the ability of soil to transmit water, which is
governed by the amount and interconnection of void spaces in the saturated zone. These voids
may occur as a consequence of inter-granular porosity, fracturing, bedding planes, or
macropores. In general, high hydraulic conductivities are associated with high rates of pollutant
transport. We use a value for saturated hydraulic conductivity (0.61 m/hr) based on the 95th
percentile of a probability distribution for hydraulic conductivity in sand derived by Carsel and
Parrish (1988). This value thus represents a conservative of "reasonable worst case" estimate.
5-23
-------
Unsaturated Hydraulic Conductivity of Soil
«
In the unsaturated zone, tide hydraulic conductivity, which is based on the effective
permeability of soil and fluid properties, is a function of the moisture content, which is in turn
a function of the pressure head, These relationships are central to the simulation of water flow
through the unsaturated zone. As inputs, the VAJDOFT model accepts sets of data points
describing effective penneability-iaturation curves and the saturation-pressure head curves
Alternatively, it accepts van Genuchten water retention parameters defining the curves (U S
EPA, 1989b; Carsel and Parrish, 1988); we select this latter option for this analysis.
Based on soils data from the Soil Conservation Survey (SCS), Carsel and Parrish derived
distributions for the three parameters required (0r, a, and 0) according to twelve SCS textural
classifications (Carsel and Parrish, 1988). Values used for our calculations (0.045, 14.5 nr1,
and 2.68 for 0r, a, and ft, respectively) correspond to the values reported for sand.
Fraction of Organic Carbon in SoU or Sludge
The model combines the fraction of organic carbon in the soil with each contaminant's
organic carbon partition coefficient to determine the partitioning of contaminant between soil and
water. In general, a lower fraction of organic carbon implies greater mobility for organic
contaminants. The organic carbon content for sludge varies among sludge types, with mean
values for various types showing a relatively narrow range of 27.6-32.6 percent (U.S. EPA,
1983b). We conservatively assume that soft within the upper 15 cm of the soil column contains
1 percent organic carbon. We seksct a value of Itf3 for the fraction of organic carbon in the
unsaturated zone because it is a typical value for sand, and falls at the lower end of the range
(0.001-0.01) reported for soil beneath hazardous waste disposal facilities (U.S. EPA, 1986f).
The fraction of organic carbon in the saturated zone is expected to be lower than that of the
unsaturated zone, and has been assigned a value of 1O4, or one tenth the fraction we assume for
the unsaturated zone.
Depth to Groundwater
The depth to groundwater is defined as the distance from the bottom of the mixing zone
to the water table. The water table is itself defined as the subsurface boundary between the
unsaturated zone (where the pore spaces contain both water and air) and the saturated zone
(where the pore spaces contain water only). It may be present in any type of medium and may
be either permanent or seasonal. The depth to groundwater determines the distance a
contaminant must travel before reaching the aquifer, and affects the attenuation of contaminant
concentration during vertical transport. As this depth increases, attenuation also tends to
increase, thus reducing potential pollution of the groundwater. We use a conservative value of
1 m for the distance between treated soil and groundwater.
5-24
-------
5.2.3 Hydrologic Parameters .
Key hydrologic parameters iinclude net recharge or seepage, the thickness of the aquifer,
and the hydraulic gradient. Values used for these calculations were included in Table 5-4 and
are discussed below.
Net Recharge
The primary source of most groundwater is precipitation, which passes through the
ground surface and percolates to the water table. Net lecharge is the volume of water reaching
the water table per unit area of land, and determines the quantity of water available for
transporting contaminants vertically to the water table and laterally within the aquifer. The
greater the recharge rate, the greater the rate of contaminant transport, up to the point at which
the amount of recharge is large enough to dilute the contaminant. Beyond that point, the two
effects may cancel each other out U.S. EPA (1985b).
For land application sites, the selected recharge rate (0.5 m/yr) represents the average
of a range of values presented in (U.S. EPA, 1986f).
Thickness of Aquifer
Saturated zones are considered to be aquifers if they can transmit significant volumes of
water. Only aquifers are considered when selecting input parameters for these calculations. For
estimating aggregate risks, we conservatively assume the thickness of the aquifer is 1 m.
Hydraulic Gradient
The hydraulic gradient is a function of local geology, groundwater recharge volumes and
locations, and the influence of withdrawals (e.g., well fields). It is also very likely to be
indirectly related to properties of porous media. Rarely are steep gradients associated with very
high conductivities. No functional relationship exists, however, to express this relationship.
For these calculations we select a value of 0.005 m/m or 0.5 percent for the hydraulic
gradient, based on an average value for groundwaters surveyed for the Hazardous Waste
Management System Land Disposal Restrictions Regulation (U.S. EPA, 1986f).
Total Suspended Solids
Calculating the amount of contaminant partitioning to solid phases in the stream requires
a value for the concentration of suspended solids within;the stream. Raw data for total
suspended solids in streams and rivers in the U.S. were obtained from the EPA's STORET
database, under the field "Total Resiidue". The data consist of annual mean total residues for
the U.S. for the years 1903 through 1991. The geometric mean of these annual values is
calculated as 16.2 mg/l and the median as 16.41 mg/L We use a rounded value of 16 mg/7
for the suspended solids content of ithe surface water.
5-25
-------
5.2.4 Chemical-Specific Parameters
*
Distribution Coefficients
Contaminant transport in soil systems is influenced by interactions between the
contaminant and soil. The affinity of contaminants for soil particles may result from ion
exchange in clay partic es, electrostatic forces between contaminants and charged particles and
mteractions with organic carbon. When all exchange and interaction sites in a soUare filed
soluble contaminants will move through the soil at the same velocity as the bulk leachate The
affinity between a soil and a contaminant is characterized by the distribution coefficient '(KD)
Representative KD values (in //kg or mVkg) are defined as the equilibrium ratio offoe
contaminant concentration in soil (mg/kg) to that in associated water (mg/J or mg/m3). Values
Zl ' ^ T^!f. ^^ * Table 5'5 and discussed Mow. Note that the organic
contaminants mtrosodimethylamine and toxaphene are not being considered in this anSyL
because they were never detected in the analytic survey of the NSSS.
mrffi .F™*2?°PhoWc °^c contaminants, KD is calculated from a contaminant's partition
coefficient between organic carbon and water: .
KD =
where:
KD = equilibrium partition coefficient for contaminant (nrVkg),
KOC = organic carbon partition coefficient (m3/kg); and
f« = fraction of soil consisting of organic matter.
The organic carbon partition coefficient for a contaminant can be estimated from its
octanol-water partition coefficient, which can be measured in laboratory experiments. Values
ofJTOC used in this analysis are shown in Table 5-6 and are calculated from the following
regression equation by Hassett et al. (1983):
log(KOC) = 0.0884 + 0.909 log(KOW)
where:
KOW = octanol-water partition coefficient for contaminant.
?2S.2S^Pd^0f PCBS' ^ KOWv*ues used for this ^ysis have been obtained from the
U S]^ " GiapMcal ^P08"1* Modelling Systems (GEMS and PCGEMS),
Polychlorinated biphenyls .(PCBs) are a class of chemicals containing 209 possible
congeners. The most common PCS mixture is Aroclor 1254, which is dominated by penta-
congeners, with about equal amounts of tetra- and hexa-congeners. In a well-ajed soil
contaminated with PCBs, however, Aroclor 1260, which contains more penta- and hexa-
congeners than tetra-congeners, is more representative of the PCBs found (O'Connor 1992)
because the less chlorinated congeners degrade more rapidly. In order to determine a
5-26
-------
Table 5-5
Octanol-W.ater and Organic Caibon Partition Coefficients
for Organic Contaminants
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Chlordane
DDT
Lindane
Polychlorinated biphenyls6
Trichloroethylene
* All values except for PCBs
Log of Octanol-
Water Partition
Coefficient*
2.13
6.12
5.11
5.54
6.38
3.61.
6.70
2.42
taken from the CHEMEST procedure of t
Organic Carbon
Partition
Coefficient1'
106
448,000
54,100
133,000
772,000
2,340
1,510,000
194
he Granhical
Exposure Modeling System (GEMS), U.S. EPA (1989d).
b KOC for organic contaminants derived from KOWv/ith Equation 6 from Chapter 15 of
Basset et al. (1983): log(KOQ= 0.0884 + 0.9091og(*0W).
c Based on Aroclor 1254, the most common PCS mixture in sewage sludge. Derived from
O'Connor (1992) and representative values from Anderson and Parker (1990).
5-27
-------
Table 5-6
•pctanol-Water Partition Coefficients
forPCBs"
Congener
2,4'
2,2',5,5'
AllPenta
2,2',4,4',5,5'
Average"
Number of Chlorines
2
4
5
6
5.5
LogKOW
5.1
6.1
6.5"
6.9
6.7
* Source: Anderson and Parker (1990).
^timated tesed on apparent linear relationship between number of chlorines on congener
e log KOW values for penta- and hexa-congeners averaged for representative log KOW.
5-28
-------
representative organic carbon partition coefficient for PCBs, we calculate an average from log
KOW coefficients (from Anderson and Parker, 1990) listed in Table 5-7. The log KOW fax the
penta-congener has been estimated to be approximately 6.5 by noting that the log KOW values
are approximately linearly related! to the number of chlorines in the congener. Averaging that
value with the hexa-congener value gives 6.7 for the log KOW. As with other organic
contaminants, the regression equation from Hassett et al, (1983) is used to convert this KOW
value to an estimate of KOC.
for organic contaminants, KD is also a function of the organic carbon content of the
sludge and soil. For the mixing .zone, we use a conservative/^ value of 0.01. The value of
10 is selected for the.4 value of the saturated zone because it is a typical value for sand. The
fK value of lO"3 is used for the unsaturated zone to reflect the higher degree of organic carbon
associated with the soil directly under the sludge application or disposal site.
For metals, separate KD values are used to describe partitioning between the water and
soil within the mixing, unsaturated, and saturated zones, and the partitioning between the water
and the sediment in the surface water. KD values used for the mixing, unsaturated and saturated
soU zones are taken from a study by Gerritse et al. (1982) and represent the results of
laboratory column tests with a sludge-amended sandy loam topsoil. The KD values we use were
listed in Table 5-5, and represent the geometric mean of the ranges provided by Gerritse et al.
(1982). For surface water, KD values are calculated according to the following regression
equation provided in U.S. EPA (1982b):
= o
t»rv
where:
KD = partition coefficient (//kg),
TSS = total suspended solids (mg/f), and
-------
Table 5-7
Distribution Coefficients for Organic and Inorganic Contaminants
Mixing Zone Unsaturated Zone
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthaiate
Chlordane
DDT
Lindane
Polychlorinated biphenyls
Trichloroethylene
Note: The distribution coefficient
(//kg)
_n
20
431
_
jy
98
^f\ *
621
330
-
OJ
1.06
4,480
541
1,330
7,720
23.4
15,100
1.94
for organic
(//kg) -
20
431
59
98
621
330
63
0.106
448
54.1
133
772
2.34
1,510
0.194
contaminants (KD) is thp.
Saturated Zone
(//kg)
20
431
59
98
621
330
63
0.0106
44.8
5.41
13.3
77.2
0.234
151
0.0194
nmHiir«f r»f rti/»
»«" o «- caAmium
layer, and 0.1% and 0.01 % in the unsaturated and
5-30
-------
Table 5-8
Statistical Parameters for Predicting the
Equilibrium Partitioning of Metals in Surface Water*
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
a
0.48x10*
4.00xl06
3.36x10*
1.04xl06
O.SlxlO6
2.91xl06
0.49xl06
ft
-0.7286
-1.1307
-0.9304
-0.7436
-0.1856
-1.1356
-0.5719
HD™
(//kg)"
63,700
174,000
255,000
132,000
185,000
125,000
100,000
' Source: U.S. EPA (1982b).
Assumes TSS=16 mg/L
5-31
-------
Henry's Law Constants
voa. " rate at Wbich ^^ Contaminants
5-32
-------
Table 5-9
Degradation Rates
Benzene
Benzo(a)-
pyrene
Bis(2-
ethylhexyl)
phthalate
Chlordane
DDT
Lindane
PCBs
Trichloro-
ethylene
Aerobic Anaerobic
Degradation Degradation
Rate (yf1? Rate (yr'1)"
16e ' Of
0.48* 0.12h
II1 . 0^
& 36J
0.04! 2.5k
1.2" 8.3a
0.063° 0.00063"
0.78" 3.3r
Unsaturated
Zone
Degradation
Rate (yr1)"
1.6
0.048
1.1
0
0.004
1.2
0.0063
0.78
Saturated
Zone
Degradation
Rate (yr1)11
0.8
0.084
0.55
18
1.3
4.8
0.0035
2.0
' Based on microbial degradation rates, except for lindane and
trichloroethylene, where hydrolysis rates are used.
b Based on microbial degradation rates.
c Estimated as 10% of aerobic biodegradation rates. Hydrolysis rates
used for Hndane and trichloroethylene.
d Estimated as the arithmetic average of the unsaturated zone
degradation rates and the anaerobic degradation rate.
8 Vaishnav and Babeu (1987).
f Horowitz et al. (1982).
g Coover and Sims (1987).
h Anaerobic rate assumed to equal 25% of aerobic rate.
' Howard et al. (1991).
j Sheltor et al. (1984).
k Castro and-Yoshida (1971).
1 Stewart and Chisholm (1971).
m Ellington et al. (1988).
0 Zhang et al. (1982).
0 Fries (1982).
p Anaerobic rate assumed to equal 1 % of aerobic rate.
q Dilling et al. (1975). :
T Bouwer and McCarthy (1983).
5-33
-------
•s ^
H .«»
!
•f
8.1
•s*
11]
•a « «?
*—- ^ .£,
*
t
T * w ' -, _
000 00
fc<* ^< ij *Tl *"*
PCGEMS(°
(atm-mVm
11
2
x
\O
to
"
a o
ate
exyl)pht
§
CO
B
Bi
et
X rH *O rt "^ O
tHlJjf}}!
-------
and no two values agree, a measunsd value is chosen in preference to an estimated one. If only
estimated, dissimilar values are available, the value most conservative for groundwater (i.e., the
lowest Henry's Law constant) is-chosen. This last circumstance occurs only for
bis(2-ethylhexyl)phthalate.
«
»
The only exception to the decision process described above is for polychlorinated
biphenyls (PCBs), which include a variety of possible congeners with different chemical
characteristics. Anderson and Parker (1990) provide a compilation of non-dimensional Henry's
Law constants for one penta-congener and three hexa-congeners. To derive a representative
Henry's Law Constant for PCBs, the three values for hexa-congeners were averaged to a single
value which was then averaged with the penta-congener value to obtain the single constant
reported in Table 5-10.
For all organic contaminants except PCBs, the dimensional estimate of Henry's Law
Constant reported in Table 5-10 has been converted to an equivalent non-dimensional constant
based on an assumed temperature of 15'C (288 K) and the following equation:
*-JL
R T
where:
T = temperature (assumed to be 288 K),
R = Universal Gas Constant (m3-atm/mol-K),
H = Henry's law constant (m3-atm/mol), and
H = Henry's Law constant (dimensionless at specified temperature).
Because Anderson and Parker report non-dimensional values for PCBs at 25°C, the average
value derived from this source has been adjusted to an equivalent non-dimensional value at
15°C.
Diffusion Coefficients
As discussed in Section 5.1, volatilization of contaminant from a land application site is
modeled with equations provided by Hwang and Falco! (1986). These equations require
estimates for the diffusivity of each contaminant in air. To Calculate diffusivity coefficients, we
use Wilke and Lee's method as described in Lyman et al., (1990). Our resulting estimates,
which are based on a temperature of 15 *C, are listed in Table 5-11.
Bioconcentration Factor and Food Chain Multiplier
Our estimates of exposure through fish consumption (surface water pathway) require an
estimate for the bioaccumulation of contaminants in fish swimming within contaminated surface
water. As discussed in Section 5.1,5, this bioaccumulation can be described as the product of
two parameters, the bioconcentration factor (BCF) and the food chain multiplier (FM). For
metals, we take values of these two parameters from U.S. EPA (1989e). For organic
contaminants, we use each contaminant's octanol water partition coefficient (from Table 5-6) to
derive the BCF and FM values listed in Table 5-12. The food chain multiplier is also a function
5-35
-------
Table 5-11
«
Diffusion Coefficients for Contaminant in Air
Difftisivity in Air
(cmVsec)"
9.06xlO-2
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate 3.27xlO-2
Chlordane
4.51xlQ'2
DDT 4.13xlO-2
4.98x10-^
Polychlorinated biphenyls 5 69xlQ-2
Tricfaloroetfaylene _ . 8.18xlQ-2
• Calcukted according to Wilke arid Lee's method, as described in'Lyman et al. (1990).
5-36
-------
Table 5-12
Bioconcentratiott Factors and Food Chain Multipliers
1
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Chlordane
DDT
Lindane
Polychlorinated biphenyls
Trichloroethylene
Bioconcentration • Food Chain b
Factor Multiplier
(//kg) (dimensionless)
350
330
130
120
., 180
100
50
7.6
11000
1700
3700
17000
110
31000
13
1
1
1
1
1
1
1
1
10
10
10
10
1
10
1
* BCF values for inorganic contaminants are taken from U.S. EPA (1989e). BCF values
for organic contaminants are derived from the following regression equation, taken from
U.S. EPA (1990b): log(BCF) = 0.79 logCK^) - 0.8. Log(KOW) values for each
contaminant are listed in Table 5-6.
b Food chain multipliers are determined from the procedure in U.S. EPA (1990b), assuming
a Trophic Level of 3 for fish.
5-37
-------
Range of
5-5.5
5.5
Food Chain Multiplier
i
10
10
log(BCF) = 0.791ogOS:CW)-0.80
Ievel
Inhalation Volume
m> of «£±££
"» *
bas.d on an inhaUUon voMme of 20
5.2.5 Size of Exposed Population
, surface water, and air pathways
1) the size of the affected jurea near each site
Each of these components is described below.
Size of Affected Area
^rese"tati0'' of «•
tion site we
5-38
-------
Figure 5-1
Zones of Human Exposure for Groundwater, Surface Water, and Air Pathways
Air and! Surface Water
Pathways
Air, S.W.,
and G.W.
Pathways
Groundwater
Gradient
5-39
-------
.
Density of Exposed Population
Number of Sites
Exposed Population
5-40
-------
Table 5-13
Sitate Population Densities
State
California
Florida
Illinois
Indiana
Michigan
Ohio
Pennsylvania
Texas
AVERAGE:
Number of Treatment
Works Facilities
4
4
6
7
5 ]
9
5 • .
5
Population density
(per mi2)*
190.8
239.6
205.6
154.6
163.6
264.9
. 265.1
64.9
194
I
* U.S. Bureau of the Census, 1991.
5-41
-------
water through the ingestion of fish iis set equal to the total population within a 3000 m radius of
the site, or the total population exposed through the air pathway. As a result, we estimate 14
rmlhon people reside within 3000 mi radius of a site and may be exposed through the air
pathway, 1.7 million people are exposed through groundwater, 7.1 million through direct
consumption of surface water as a source of drinking water, and 14 million through the ingestion
5.3 RESULTS AND DISCUSSION
5.3.1 Baseline Risks: Groundwater, Surface Water, and Air Pathways
Table 5-14 provides estimates of cancer risks from the land application of sludge through
groundwater, surface water, and air pathways. As shown by the table, we estimate that land
application of sludge under baseline conditions causes about 1 case of cancer per ten years This
total is dominated primarily by exposure to PCBs, which accounts for about 54 percent of the
total: 36 percent is from the air pathway and 18 percent is from the surface water pathway
BAP also contributes a significant amount to the total risk (about 37 percent), through the
surface water pathway. The cancer risks from both PCBs and BAP through the surface water
pathway are the result of ingestion of contaminated fish (due to the fact that they both have high
partition coefficients and low degradation rates). Interestingly, arsenic, which has the highest
% individual nsk (4x10* for the HEQ ithrough the groundwater pathway accounts for less than 6
percent of the total cancer risk. This is due to the smaller population exposed to groundwater
compared to that through the air and surface water pathways. However, actual risks are likely
to te lower than those estimated, sines these estimates axe often based on generally conservative
Exposure to non-carcinogens through the groundwater and surface water pathways is
generally quite low. Average exposure from sludge (without background) is never greater than
one tenth of one percent of the RfD (Table 5-15), and estimated exposure from sludge (without
background) for the HEI is less than 10 percent of the RfD (Table 5-16).
5.3.2 Benefits from Regulatory Controls
B> limiting annual or cumulative loadings of pollutant from land-applied sludge to
agricultural land, or by otherwise controlling management practices for land application the
regulation is likely to reduce potential exposure through these pathways. However such
reductions in risk are unlikely to exceed our estimate of current (or baseline) risk For this
reason, we estimate that the likely health benefit for groundwater, surface water and air
pathways for this management practkie is less than one case of cancer or disease from lead per
hundred years.
5-42
-------
Table 5-14
Baseline Cancer Cases1
Land Application: Groundwater, Surface Water, and Air Pathways
AGGREGATE RISK
Arsenic
Benzene
Benzo(a)pyrene
Bis(2)ethylhexylphlate
Chloidane
DDT
Lindane
PCBs
Trichloroethylene
Total
AVERAGE INDIVIDUAL RISK6
INDIVIDUAL RISK FOR HET1
Groundwater
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
IxlQ-*
4xlO-5
Vapor
; ' —
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.04
<0.01
0.04
6xlO-7
5x10*
Surface
Water
<0.01
<0.01
0.04
<0.01
<0.01
<0.01
<0.01
0.02
<0.01
0.06
3xlOl7
5x10-*
Total"
<0.01
<0.01
0.04
<0.01
<0.01
<0.01
<0.01
0.06
<0.01
0.1
5xlO'7
4xlO-5
* These results are based on reasonable worst-case input parameters and assumptions.
b Individual totals may not sum to totals because of independent rounding
0 Risk for average exposed individual of developing cancer from lifetime of exposure to
pollutants from sludge.
d Risk for Highly Exposed Individual (HEI).
\
5-43
-------
so
r
O
II
U
3 §
fe
ll
•8.1
'S &
I?
o
ft,
t
1°
0 C5 0 0
V V V V
o o
V
"8
«3
«r a
0 «b
5
o
en 1-1
o\ cs
X
10
Q A 60
222222
— < es o\
g
s
il
^ 3 jfi ff
3 a
P
OT
g
fr
E!
B3
CO
,8
<4^ "**
"8 I
w 1
1 &
f -s
I
s
(90
II
s
Ui
P
*
'.S
•§
I
* 1 K'l
60 g Q^ g
•§ i§:!
— Hi' 55 (Li
ll&a
-------
O
Q
u
I
i &
o
e
•c
o.
o
o
•o
J*
n
CO
X
4
o
i—i
X
§
o
V
^ 2
en »—i
X
en
X
oo
i—<
X
tn
X
-* en •-< ?s
X
CN
X
en
x
CN
X
oo
X
CN
O
1«H
X
— en *•*
es
CN
O
E
^3
I'
•:2
(3 P»
j~ o
O U
I
I *
o
V
T
vn
.9
-------
6. LAND APPLICATION: RESIDENTIAL USES
6.0 INTRODUCTION
v2±bte ST1 r5' I""1? 8ari?°erS' Wh° *&* *" P™*«« to *eir home
Adults also ingest small quantities of sludge-amended
6.1 METHODOLOGY
6.1.1 Overview
6-1
-------
Another difference between residential and non-residential land application is that the
analysis for home gardeners pays special attention to a particular demographic group: children
of ages one to six years. These children are assumed to be exposed to additional risks from the
ingestion of sludge-treated soil. To estimate population risks from this pathway, we assume the
average child in a gardening household ingests soil at a rate of about 0.2 grams per day. This
value is based on a 1985 pilot study in which tracer elements were used to estimate the amount
of soil ingested daily by 70 children living near a lead smelter (Binder et al., 1986).
Using findings based on aluminum levels measured in soil, dust, and the children's stool
samples, we assume that 0.2 grams per day represents mean soil ingestion for all children.
Since children are likely to ingest soil from a number of locations other than their yard or garden
(e.g., schoolyard, playground), the models assume that only 10 percent of all soil ingested has
been treated with sludge. The contaminant doses ingested through this pathway are then added
to those the children receive by eating vegetables from the home garden. Our estimates for the
amount of vegetables consumed by these children are based on the amounts reported for the one
to six year age range of the Pennington (1983) study. Because exposure to this amount of soil
and these quantities of vegetables is limited to approximately 5 of the 70 years in an individual's
expected lifetime, we adjust estimated cancer risks for children by a factor of 5/70 when
estimating total lifetime risk or incremental cases per year.
Adults are also assumed to ingest some soil, but in quantities much smaller than those
ingested by children. U.S. EPA (19'88d) cites a value for adults' dust ingestion of 0.02 g/day,
or about 3x10^ g/kg-day. As with children, we assume that only 10 percent of the total amount
of soil ingested is from sludge-amended areas. Exposure through soil ingestion is added to
exposure through consumption of home-grown vegetables: We take our assumed rates of
consumption for vegetables from the Pennington (1983) study. Cancer effects are adjusted by
a factor of 65/70 to reflect the fraction of an average lifetime when soil is consumed at adult
rates.
The individual exposed to the most health risks from the use of sludge for residential
purposes (the HET) is assumed to be a young child who lives in a household using sludge on its
vegetable garden and who has a tendency to ingest soil materials at above-average rates. Such
a child is assumed to ingest 0.8 g of soil per day, or about 0.08 g/kg-day (U.S. EPA, 1989h).
Finally, this analysis models risks to those who use sludge only for ornamental
gardening, j»uch as landscaping or lawn care. These risks are incurred through ingestion of soil
by children and adults. All assumptions regarding soil ingestion are the same as those described
above. The model assumes that households that use sludge for ornamental gardening do not also
use it for home vegetable gardens, and vice versa.
6.1.2 Description of Calculations
Our calculations for exposure and risk to individuals using sludge on their home gardens
differ from those for land application in the following respeicts:
6-2
-------
(1) We assume that no meat is produced from crops grown in home gardens.
(2> ^^L^Tte11™ ^lt'pToducf ft?01 home gardens is distributed to national
(3) Children of home gardeners are assumed to ingest a certain average amount of
garden soil daily, m addition, adults are assumed to" ingest somf gaXLS
daily, but in lesser quantities than children.
ld t ; 5T d?dren throu«h direct ^estion of soil, we assume
children ingest 0.02 g/kg-day (average) or 0.08 g/kg-day (for the highly exposed
individual) of soil. Our estimates of cancer risk tough this pJto!*™
adjusted by a factor of 5/70 to reflect the estimated five-year duration of tne £5
mgestion behavior. We assume that adulte ingest soil at a rate of 0.0003 e/kg-
day, and adjust their estimated cancer risk from soil ingestion by a factor of
f™ £*** ^ te and cMdren' exP°s
exposure from ingestiom of garden produce.
and cMdren' exP°sure om <"** «"» is ad to
(5> Jvi^ate lMa fr?m «*niuift- exposure, estimated cadmium burden from
t0 eStimatWl bUlden *»•-< exposure when
*.
For households with vegetable gardens, the calculations proceed in three steps First
«n£T? ^r^118 °f Sludge TOnstitu^ ta the tissues of p'
gardens. Second, we calculate individual and population risks based on
on
produce. Finally we sum estimated risks from aU crops and soil mgestionto
from each «,ntaminant. For ornamental gardeners, ^calculate risSC soo
M cakuktions are repeated with and without application of sludge, so that ESS^
by background so* concentrations or inlake from other sources arlnot attributed to skTge
calculations used for each component of the analysis are described below.
Concentratfo^ «f Contaminants ^ in Produce from Sludge-Amended
ftnm H ff 4 Cal(:,Ulate ^ maSS °f '""* «««niiiaiit in die niixing zone at background levels
from the background concentration of each contaminant and the masl of soil in theSxing zon*
LB. = MSH BS. 10"3
where:
= mass of contaminant J in mixing zone as a result of background
concentrations in ssoil (kg/ha),
MSH = mass of soil in mixing zone (Mg/ha),
6-3
-------
BSj - background concentration (dry wt) of pollutant 7' in soil (mg/kg or g/Mg),
and ' __-~
10'3 = constant to convert units from (g/ha) to (kg/ha). i- \
* ' ".*„_ ' ."'
We calculate the mass of contaminant added through the application of sludge as:
LAj = N Cj AR 1(T3
where:
LA, = mass of contaminant j added by application of sludge to home garden /
(kg/ha),
N = number of years sludge is applied to home garden (yr),
Cj = concentration of pollutant j in sludge (mg/kg or g/Mg),
AR = application rate of sludge (dry Mg/ha), and
10"3 = constant to convert units from (g/ha) to (kg/ha).
We sum these two estimates and divide by the mass of soil in the mixing zone (adjusted
for additional soil mass from sludge) to approximate the concentration of contaminant in treated
soil:
I00°
(NAR) + MSH
where:
CTj = concentration of contaminant/' in garden soil, adjusted for background soil
concentration and for additional soil mass from added sludge (mg/kg or
g/mg), and
1000 = constant to convert units from (kg/Mg) to (mg/kg).
Note that as N approaches infinity, C2} approaches Cj, so the application of sludge cannot
increase estimated concentrations of contaminant in the soil beyond the concentration in the
sludge.
Ohce we have estimated the concentration of contaminant in soil, we use estimated uptake
rates to calculate the expected concentration of contaminant in crops grown on that soil:
where:
= tissue concentration (dry wt) of pollutant/' in crop / (mg/kg or g/mg), and
= rate of uptake of pollutant j into tissue of crop i (mg/kg dry weight per
mg/kg in soil).
6-4
-------
flrmoJ^ T^ ^ • 6X?°SUre to contaminants ft™* sludge, we combine these estimated
H * TT of,confanufant m SK*™ P«*"<* with assumptions about dietary consumption
and the fraction of national produce grown in sludge-amended land. We also consider additional
exposure that may occur through the inadvertent ingestion of small quantities of sludge by
children or adults: » ° J
EXP, = £CD, FC. DC. 10-' * *"«A*C,W* (1-lM) /, AR C, KT*
/. .. ' 'AHZT M& -
where:
EXP,. = total exposure of pollutant./ from fruits, or vegetables grown in the home
garden (mg/kg-day),
FQ = fraction of dietary consumption of vegetable or fruit / grown in home
garden (dimensi:onless),
DQ = daily dietary consumption of vegetable or fruit (g/kg-day),
10' — constant to convert units from (g) to (kg),
DA = duration adjustment for childhood exposure through direct ingestion of soil
(dimensionless),
IK = average rate of soil ingestion for children (g/kg-day),
LS = average human llifespan (yr), and
I» = average rate of isoil ingestion for adults (g/kg-day).
*.
Detennming ^dividual and Populatiion Risks from Contaminant Ingestion Through Foods
or Soil from Sludge-Amended Gardens
Cancer Risks
For contaminants classified as carcinogens, we convert this estimate of exposure into an
incremental nsk of cancer for the exposed individual:
where:
^ =~ uppers-bound-, estimate of individual cancer risk for contaminant j
(probability of developing cancer from lifetime dose of EXPj), and
Qj = human cancer potency of contaminant j (mg/kg-day)"1.
We assume that small incremental risks from individual contaminants are additive, and
sum to calculate the total incremental risk for exposed individuals:
where:
6-5
-------
CI = total individual cancer risk (incremental risk of developing cancer within
lifetime as result of land application of sludge).
*
We combine this result with an estimate of the number of home gardens using sludge,
to calculate the incremental number of cancer cases caused annually by the use of sludge in
home gardens:
LS
where:
CP = total aggregate risk of cancer from the use of sludge on home gardens
(number of additional cancer cases expected per year), and
POP = exposed population.
Health Risks from Non-carcinogens
For non-carcinogenic contaminants, we express risk as the ratio of expected exposure to
the risk reference dose (RfD) for each contaminant of cpricern:
= POP (IF EXPj+BIj >
= 0 (IF
wjiere;
NCPj = number of people exceeding RfD of pollutant j due to exposure from
sludge,
BIj = background intake of pollutant j (mg/kg-day), and
RfDj = Risk Reference Dose for pollutant/ (mg/kg-day).
This calculation is repeated with and without the application of sludge; the difference in NCPj
for these two scenarios provides a measure of aggregate risk.
The methods we use for estimating non-cancer health effects from lead and cadmium
differ from those used for other non-carcinogenic contaminants, as explained in Chapter 2.
Lead's effects are estimated separately for men, women, and children. The model bases
calculations on estimated blood lead levels, and uses nonlinear dose-response functions to
estimate non-cancer effects. For cadmium, expected cases of kidney disease are estimated
separately for smokers and nonsmokers. Calculations are based on kidney cadmium uptake
rates, and exceedence of a threshold of 200 /tg/g cadmium in the kidney is defined as a "case".
These methods do not imply that everyone exceeding this threshold will experience kidney
disease; rather, those exceeding this threshold are considered to be at risk from this health
hazard.
6-6
-------
6.2 DATA SOURCES AND MODEL INPUTS
innntc °f """I?** ^^ ti^ ^^ residential uses of sludge relies on a number of
inputs and assumptions. The discussion below describes the inputs to the risk assessment mode?
as well as the assumptions and data from which the inputs were derived. Table 6-1 sunTaSes
S T^ Pf1^61615 «? d ^Ptions. Although we use the best available estimate for each
input in the baseline analysis, the true value of many inputs is uncertain.
6.2.1 Volume of Sludge to Home Gardens
According to the National Sewage Sludge Survey, about 125,000 dry metric tons of
987^ SJd^;?17.? *4 «« f*>«*l »** annually. The National oLenSg £££
¥,**? °n °f *" 69 mmion househ°lds Evolved in gardening activity in 1986
We ^{OTt aSSUme ** ab°Ut ^ of ^ l««*Sd. using sludge apply it
^
to *™» « oniamntat sband
quantities °f
6.2.2 Application Rates
^tes estiniated for agricultuial uses of sludge, we assume the home
,
sludge per hectare per. year to horher home vegetable
.
aPPHcation rates but held constant the quantity of
6 a smaUer ^timate for the number of households
*erel°re relatively ^--Wve to application rate assumptions a
f ^ u^6 "° to8110141' b«t non-cancer risks would be affected by
Kn^S> ^ eSVer W11^0^ to any particular garden increase the
household's probability of exceeding RfD thresholds for one or more contaminants indulge.
6.2.3 Concentrations of Contaminants in Sludge
h**™,™? M~2 UStS m^ *** 99th P6^04116 concentrations of sludge constituents
background soil concentrations for these contaminants, and adult and child background hSke
-
contaminants do not accumulate in soil with repeaapplications
l° ^S*16 levels b^een yearly harvests. MeSs do 2t decay;
be°onie wnavaaable «»• Pto uP^e, or be carried away from the
re ah ? ^g '^^ H°Wever' field studies «*** that metals can
remain available for plant uptake for several years (see, for example, Heckman et aL, 1987).
For this analysis we assume that sludge is applied for 20 consecutive years and tha metals
ve years an a metas
accumulate and remain bioavailable. Siace this analysis is concerned primarily with inTrenS
cancer risks from the land application of wastewater sludge, background soil concentrations and
mtakes of carcinogens are not of particular concern for the estimation of cancer risks. They are
important in the estimation of non-cancer risks, however, since non-cancer risks are assumed
6-7
-------
•a
0
<- !2
• r
c .a
.2-3,
t£
Assumption
1
cu
li
{i
"
§
*
J3
I
«
i
is
II
S ss
09
•o
o
S3 "S
*l
3 33
z <
i
I
o
u<
4>
" S
|
O
g
03
g o
•O M
8 I
-------
I 1
w Hn 1 1
'-N S !jg
*§ 1 1 •
a a* o \
•£ hS
elS
O F~ (^ 1 1
<§ P - „
. -a a
^ S .2
ve rt -.3 j
« co
1 f 1 a
£^ M4 ~7* 1 »S5 1
12 5
"1 I '
t <
2
2
<
1
,0
4— »
u
e
cS
•S
ca
u<
I
i
E
S^
3
_c
I
^™^
flj
5
1
1 1 W
'S
ca
CO
J
O
£
^
*+-t
• »H
1
1
*-»
^
4)
1
»-
O
h
II
s
.s3
1
1
•s
1
s
11
1}
C bD
w ^^
II
t
A4
.w
'C
i
i
£: .
1
t*
0
A
1
1
•s
00
!
C
o
•a
1
'3 ™'
•a o
«i
18
1 0 percent of t(
e-amended gard
>»Wl
31
* * «a ' 5
8-S -e
j= «- =•
S 8 s
s § f
S 2 fe
11 o
S "" *
1 1 i
8 S S
« •- s
s "a **
^3 jUj ^^
i ^2 -a
II !
Co t-c
.2? " «a
if If
S3 » -S-S
D£ Kl
., 11
I !*
1 II
^+-« -_^ CQ
J IB
I 11
« .2 S
1 11
s .
"H S^
i «i
Sf 3 § ti
S S § S
JC tf"t *O OHM
co O O to
•S a «2 a
II III
4)
w.
1;
o
2
cd
w
on-cancer
ifetime.
G ^"^
di .OT
•§ g
1 &
5 W
|.s
^.s
s a
4-»
§
1
U< ;
1
"e9
> d
e o
•si
tl
lii
n:
13 -s
«} O,
5P g
CQ H<
4> 52
»> C3
•< 8
C
o
"•J-3
r*
1
6
-------
to have thresholds (risk reference doses or RfDs) below which no adverse health effect is
expected. Background concentrations and background intakes must be known to determine
whether an individual's total exposure to a contaminant exceeds the RfD when sludge is added
to the soil.
6.2.4 Uptake Rates for Crops
Table 6-3 lists the uptake rates for sludge constituents into the tissues of vegetables grown
in a typical home garden. The values represent geometric means of the plant response curves
from field experiments for sludge. All rates are reported in units of mg/kg dry weight of plant
tissue per mg/kg of dry soil.
6.2.5 Dietary Assumptions
The quantity and types of food! consumed by those with home gardens are assumed to be
the same as for those without gardens. Table 6-4 lists our assumed dietary consumption rates
(in g/kg-day) for both the average adult individual and for a one to six year old child (U.S.
EPA, 1989h). The values for adult consumption have been obtained by averaging consumption
rates across the sexes in the 14-16, 25-30, and 60-65 age categories, and then calculating an
estimated lifetime rate based on all age categories. Table 6-4 also lists assumed values for the
fraction of each food consumed by a gardening household that has been grown in that
household's garden (i.e., that has been produced in sludge-amended soil). These fractions are
abbreviated as FC, and represent consumption patterns for a rural, non-farm resident (U.S. EPA,
1992c).
6.2.6 Exposed Population
Table 6.5 describes the calculations used to estimate the number of individuals exposed
through dietary consumption of foods grown with residential use sludge. From the 1986-1987
National Gardening Survey, the average garden size for combined urban and rural home
vegetable gardens is 1,670 square feet, or 0.016 hectares. Three additional assumptions allow
estimation of the number of individuals exposed to risks from residential use of sludge: we
assume that 125,058 dry metric tons of sludge are applied to residential land annually, that the
sludge is applied at an average rate of 11 metric tons per hectare, and that the average gardening
household includes an average of 2.7 persons. With these assumptions, we estimate that about
1,900,000 persons are potentially exposed to health risks from residential uses of sludge.
To estimate the population of children exposed through the soil ingestion pathway, we
assume the proportion of children in gardening households is the same as the national average
(approximately 7.5 percent). As discussed earlier, we also assume these children ingest soil at
the median rate estimated for all childlren, and 10 percent of the soil ingested is composed of
sludge-amended soil. With these assumptions, we estimate the children's exposure and health
risks from contaminants in the sludge.
6-10
-------
Table 6-2
Pollutant Concentrations in Sludge and Soil
Aldrin/dieldrin
Arsenic
Benzo(a)pyrene
Cadmium
Chlordane
Copper
DDT/DDE/DDD
Fluoride
'--
Heptachlor
Hexachlorobenzene
Hexachlorobutadiene ""
Iron
Lead
Lindane
Mercury
Molybdenum
Nickel
PCBs
Selenium
Toxaphene
Zinc
Mean
Concentration
in Sludge
(rag/kg)-
0.021
10
11
10
0.25
520
0.021
0.020
11
•-'-'—<)" '" "
n
\J
140
0.025
3.6
11
66
1.3
.,7.4
0.99
" 1,300
99th Percentile
Concentration
in Sludge
(rag/kg)"
0.10
62
67
120
1.3
2,500
0.12
0.10
667
0
490
0.13
18
51
980
6.1
49
5.1
33,000
Background
Concentration
in Soil
(mg/kg)b
0
3
0
0.2
0
19
0
0
0
0
0
11
0
0.1
2
18
,0
0.21
0
54
Background
Intake for
Adult
(mg/day)
0
0.082
0
0.027
0
0.16
0
0
0
0
0
0
0.11
0
0.0066
o
0.17
0
,0.12
0
13
Background
Intake for
Child
(nig/day)
0
0.041
0
0.016
0
0.074
0
0
0
0
0
0
0.067
0
0.0066
0
0.0096
0
0.059
0
6.7
* Mean and 99th percentile sludge constituent concentrations obtained from the
National Sewage Sludge Survey.
b Background soil concentrations from U.S. EPA (1988d).
analytic survey of the
6-11
-------
fi
vi
1
a
4>
"O
O
§
6
o
SI
o
tj
5
:=»
^
a>
(90
g
55
i
l
22
H
1
o
us
4J
^
JO,
§1
& cj
1
<£
^» w
tj o
i!
- S
f*2
1
^
11
S 2
l'§
6^
w
°&!
0
O
o
o
o
o
o
a
c
•a
"u
Aldrin/Di
ws
d
8
d
^^
8
d
OO
*-4
o
d
8
d
8
d
8
d
Arsenic
o
o
o
o
o
o
o
D
S
:>,
A
"a?
«
00
S o
d
8 o
d
VO
8 e
d
^O
»-< ^
d
1 °
d
So
d
\o
§ o
d
Cadmium
Chlordane
S3
g 0 0 0 0 0
d
800000
d
^
§00000
d
_«
o o o o o o
d
g O O O 0 O
d :
00
o o o o o o
d
g o o o o o
d
.. ., ., ..
§ .1
J3
J
[14 b, O O
Q 0 .§ J S
j- Q -a •§ 3 3
a, f: '= s § 8
§- 5 § 8- g g
U Q E K K 32
o go
d
080
d
C"4
o 8 o
d
\£)
o § ®
d
080
d
080.
d
JS
0 g ©
d
o
1 J .1
i ;°. i o s o
odd
S 0 g 0 g 0
d 0- o
-< °° 00
8 *° s o S o
d <=> S d
oo •* ts
I o S o g o
d : d d
00 W1 ^
0 0 8 O g O
d ., d 0-
^f \Q \O
S o § o S o
did d
*-> ;_ 33 oo
8 « § o S o
d <=> § d
g
3
rJL ii
§ S 2 a •§ §-
5 o o ri JS ><
^r" *^ V£ w o O
^ ,2; Z O-i CO H
00
d
^-t
0
d
t"*
o
d
Os
2
d
SO
i
d
00
d
o
d
1
.
^
CO
•1"
•8
1
o
•s
1
"B.
o
f
60
^J
B*
i
•a
>
i
OS
<
S
HM
CO'
*-^
1
o
cs
-------
Table 6-4
Dietary Assumptions for Land Application:
Residential Uses
Dried Legumes
Garden Fruits
Sweet Cora
Leafy Vegetables
Non-Dried Legumes
Potatoes
Root Crops
Fraction of
Consumption from
Sludge-Amended Soil*
0.59
0.59
0.59
0.59
0.59
0.37
0.59
. — ____
Average Consumption
for Adults
(g/kg/day)b
0.036
0.059
0.10
0.028
0.089
0.22
0.023
.
Average Consumption
for Children
(g/kg/day)b
0.13
0.17
0.40
0.049
0.33
1.0
0.067
* From U.S. EPA (1992c).
for
6-13
-------
Table 6-5
Estimated Population Affected By Residential Uses of Sludge
Size of Average Garden (ha) 0.016*
Application Rate (DMT/ha) 11
Sludge to Home Uses (DMT) 125,058b
Persons per Plot 2.7"
Persons Exposed ' 1,900,000
* From 1986-87 National Garden Survey (1987).
b From the Analytical Survey poition of the National Sewage Sludge Survey (U.S.
EPA, 1989a).
c Average persons per household. U.S. Bureau of the Census, Statistical Abstracts of
the United States: 1992. 112th Ed.. Washington, DC., 1992.
6-14
-------
6.3 RESULTS AND DISCUSSION
6.3.1 Baseline Risks
^toateftha?1 CanCCr ***** f°r res!dentM uses of sludSe are presented i
Because uptake of
n
greater ton Ae Rffi, but inc^^j exposure from ^ .„
Table 6.8 compares estimated exposure to the RiD for the hiehlv exnose
To esumate the HH's exposure, we combine "reasonable worst cas? nSdSSm
"
tions ' exposure
that ab2e°WSeSt™at^h^risfe^m^^^^ Weestimate
5" * threSh°ld ^
a sof i in their kidneys as
fadMdu»Ia (°f which about 70 P6 are
ta ^ A —- — r (about 500)
6-15
-------
Table 6-6
Baseline Cancer Risks for
Land Application: Residential Uses*
Cancer Risk
AGGREGATE RISKS"
Aldrin/Dieldrin 4x10*
Arsenic 0 01
Benzo(a)pyrene 0.001
Chlordane 3x10"*
DDT/DDE/DDD 8x10^
Heptachlor 1x10*
Hexachlorobenzene 2xlO~*
Hexachlorobutadiene 0
Lindane 4xlO'7
PCBs ixio*
Toxaphene IxlO"5
Total 0.01
AVERAGE INDIVIDUAL RISK6
INDIVIDUAL RISK FOR HEl[d 1x10^
* Approximate size of exposed population: 1,900,000.
b All values in incremental number of cancer cases expected per year as a result of
exposure through residential uses of sewage sludge.
0 Risk for average exposed individual of developing cancer irom lifetime exposure to
pollutants in sludge.
d Risk for the Highly Exposed Indivdual (HEI) of developing cancer from lifetime of
exposure to pollutants in sludge.
6-16
-------
-S
o
£
V2
a .
g
M .13
c« <->
'II
g
a
u
I
3 S E i 11 I I
-------
2
s
3 a
. s
*O w
S 4)
s s
a. 8
Ex
^1
S.J
o a,
.-o
8 §
rrl co
^^ ^}
O •»
(U
a
CO Ql
s
a
o
1-4
60
"S-8"3>
« S S
"If
•?
| 1st
w «
-HH BO
S ^ 00
JS «n oo
X
Arsenic
CD irj
CO ON
^
'S
oo
o o o
m ?n en o
«n ^H ^H
en
o <=>
-------
Table 6-9
Baseline Noncancer Health Risks for
Land Application: Residential Uses'
• Health Risk
CADMIUM (Persons Crossing Kidney Cadmium Threshold
Smokers
Non-Smokers
Total
LEAD (Persons Crossing Blood Lead Thresholds)
Women
Children
Total
LEAD (Estimated Cases/Yr)
Children
°f
900
20Q
1000
40
5QO
500
to
6-19
-------
6.3.2 Benefits from Regulatory Controls
Regulation of residential uses for land applied sludge might reduce these baseline risks
by eliminating residential uses for sludge with relatively high concentrations of pollutants, or by
otherwise altering current practices. The extent of this benefit could not be determined from
existing information, but the benefit is unlikely to exceed our estimates of risks under current
(baseline) conditions. If the regulation eliminated ajl of the existing health risk, it would provide
a benefit equal to the avoidance of about 2 cancer cases per 100 years and about 500 cases of
non-cancer disease per year.
6-20
-------
7. SURFACE DISPOSAL
7.0 INTRODUCTION '
The surface disposal of municipal sewage sludge can cause pollution of groundwater and
thus nsk to humans who drink the groundwater. Emission of volatile organic pollutants from
surface disposal units can also result in exposure and risk for humans who inhale contaminated
not considered for this analysis.
,u *• ^e,U:S;^Atodefinedth« term "surface disposal" broadly: the definition includes
the disposal of sludge in waste piles, llagoons, sludge-only monofills, dedicated land-application
sites, and other practices. Two idealized prototypical facilities have been defined to represent
this diverse mix of related management practices: a monofffi and a surface impoundment.
For the monofill prototype, the facility is represented as a sludge-only trench fill
Disposal involves the excavation of trenches in which the sludge is entirely buried below the
original ground surface; de-watered sludge may be directly deposited in the trenches from a haul
vehicle. Only de-watered sludges with solids content greater than or equal to 20% are assumed
to be suitable for disposal and the sludge is often mixed with a bulking agent (e g soil) to
increase solids content. Normal operating procedures require daily coverage which reduces
odors and provides vector control, with a final cover placed on the monofill after closure.
In the surface impoundment prototype, the facility is assumed to receive a continuous
inflow of low solids wastewater sludge. A vertical outflow pipe maintains the surface liquid
level at a constant height, and liquid is assumed to leave the impoundment both in the outflow
and in seepage through the floor of (lie impoundment. Sludge entering the impoundment is
assumed to have a low solids content: (between 2-5 percent); Over time, particulate settling
occurs and a denser layer of sediment accumulates on the floor of the lagoon. Eventually this
layer of sediment reaches the top of the impoundment and no further inflow is possible. Upon
closure, the sludge is left permanently in place and remains uncovered.
One key difference between the surface impoundment and monofill prototypes is that the
active surface impoundment is assumed to contain significantly more liquid than the active
rnonofdl. Seepage through the floor of the facility is therefore expected to be greater for a
surface impoundment, and may be sufficient to sustain a local "mounding" of the underlying
water table. The surface layer of the impoundment is also assumed to be in a liquid state over
the active lifetime of the facility. The volatilization of organic contaminants from this liquid
layer is expected to differ from that predicted for a monofill, which is assumed to contain a
higher percentage of solids and to receive a daily and eventually a permanent soil cover.
7-1
-------
As discussed in Chapters 1-6, our analysis uses the sample of facilities in the analytic
survey of the NSSS to represent the complete national inventory of treatment works. For
surface disposal, we have assigned each POTW reporting the use of surface disposal in the
analytic component of the NSSS to either the monofill or surface impoundment categories.
Those reporting the use of "dedicated" or "monofill" surface disposal are represented by the
monofill prototype; those reporting the use of "other" types of surface disposal are represented
by the surface impoundment prototype. In addition, some facilities originally reporting the use
of land application have been reassigned to surface disposal if they appear to be applying the
sludge at high rates to dedicated sites; these reassigned sites are modeled using the monofill
prototype of land application.
Based on the modeling of human exposure and risk for those treatment works believed
to practice surface disposal, estimates of individual and population risks are prepared for each
site. Results are scaled to the national level with sample weights from the NSSS. Exposure and
risk are estimated both before and after the regulation. Based on several "worst case"
assumptions, we provide conservative estimates of likely, risks under current or "baseline"
conditions. Because most facilities are expected to comply with the numerical criteria even
under baseline conditions, the actual reduction in health risks to be achieved by the regulation
is expected to be significantly smaller than the predicted baseline estimates.
7.1 METHODOLOGY
As with groundwater and air pathways of exposure from land application (Chapter 5),
our general strategy for evaluating these pathways is first to determine the expected behavior of
organic and inorganic contaminants loaded into the surface disposal facility. We begin by
estimating the fraction of contaminant likely to be lost through volatilization, leaching, and
chemical degradation. These calculations, which we refer to as "mass balance," are based on
the principle that contaminant mass is conserved; the total mass of sludge contaminant lost to
these processes or retained in the soil cannot exceed the total loading.1 Because the physical
characteristics of the two modeled prototypes are different (one contains uncovered liquid, the
other a covered mix of sludge and soil), we use different mathematical models to represent
losses of contaminant for each prototype.
After completing the mass balance calculations, we use additional mathematical models
to p^-edict the movement of sludge contaminants through environmental media. We then
combine our results with data describing the densities of human populations to estimate likely
human exposure and risk. Details of each of these steps are provided below.
'For this analysis we ignore the possibility that one contaminant may degrade into another
contaminant.
7-2
-------
7.1.1 Algorithms for the Monofil! Prototype
Methodology for Mass Balance
Contaminant mass is assumed to enter the facility through daily deposits of sludee and
J±3T£ ?™& d^tio«, leaching and volatilization. B^ rflLnS^ ±
assumed to be first-order (that is, proportional to the residual concentration of
to teSflTlS'r88 ^"^ Leachin8- A client for the rate of contaminant loss
to leaching is calculated by assuming that contaminant mass in a filled monofill cell is Zoned
l™±"UmA***T tS°!.Ved md ads0lbed P"3**' Based on mathemati^ SS
2SS? m £^ndlX Jh.?6 oonccntt»ti« ^ contaminant dissolved in water wS the
monofill can be estimated from the total concentration of contaminant within thefecmt^
where:
and:
Cw - concentration of contaminant in water-filled pore space of sludge/soil
(Kg/m ),
C, = total concentration of contaminant in sludge/soil (kg/m3)
BD = bulk density of sludge/soil (kg/m3), "
H = Henry's Law constant for the contaminant (atm-m3/mol)
H - Henry's Law arastant for the contaminant (dimensionles's),
- equilibnum partition coefficient for the contaminant (m3/kg)
= air-filled porosity of sludge/soil, (dimensionless),
- water-filled porosity of sludge/soil, (dimensionless)
= ideal gas constant (8.21xlQ-5 m3-ato/mol-K), and
= temperature (K).
For anarbiteary unit concentration of contaminant in the sludge/soil (1 for/m3) a flux of
'
conn Cuonce^tionofcontaminantinleachate. Moreover, withaunk
concentration of contaminant, the mass of contaminant beneath one square meter of surface is
equal to the volume of sludge/soil beneath that area. This
7-3
-------
contaminant/m2 area/m depth of the mondfill. As discussed in Appendix B, the estimated flux
of leached contaminant is divided by this mass to derive a first-order loss coefficient for
leaching:
NR
where:
= loss rate coefficient for leaching from monofill (yr1),
NR = annual recharge to groundwater beneath the monofill (m/yr), and
= depth of a monofill cell (m).
Contaminant Losses to Volatilization. Rates of volatilization from a filled cell in a
sludge monofill will vary according to whether or not a cover layer of soil has been applied.
We assume that each cell in the monofill contains uncovered wastewater sludge for a few hours
on each of the days it receives sludge. Following each deposit, a temporary cover layer of soil
is applied. Once the monofill' s capacity is exhausted, a thicker permanent cover of soil is
applied to the entire facility (U.S. EPA, 1986h). A time-weighted average of emission rates
with and without cover is therefore used to describe the average rate of volatile emissions for
an individual cell in the monofill. Ilie fraction of the facility's active lifetime that a typical cell
will be uncovered is calculated as:
Jun LF
where:
fm = fraction of facility's active lifetime!that a typical cell contains sludge
without soil cover,
tun = time that a typical monofill cell contains uncovered sludge (yr), and
LF = active lifetime of monofill (yr).
Some monofill cells will be filled early in the facility's operation, others closer to the
facility's closure. We assume that the average monofill fell will contain sludge for half the
active lifetime of the facility. The fraction of the facility's active lifetime that such a cell will
contain siudge protected by temporary cover is:
Jco ~ "2 •'«« -...-.-• - ......
f^ = fraction of facility's active lifetime that typical cell contains sludge with
temporary soil cover (dimensionless).
A time-weighted average rate of emissions from a typical monofill cell is calculated from
equations describing emissions from a cell with and without soil cover. According to
7-4
-------
(1986h)- emissions
0.17 « 0.994(r'293> C
(7-1)
where:
= emission rate from treated soil for uncovered period (kg/nf-sec)
= wind speed (m/sec),
= temperature (K),
- concentration of contaminant in air-filled pore space of treated soil
(Kg/ or), and
= molecular weight of contaminant (g/mole).
For a cell with soil cover:
T
C.
where:
9.2x10
-*
emission rate from treated soil for covered period (kg/nf-sec)
au--fiUed porosity of cover layer of soil (dimensionless)
total porosity of cover layer of soil (dimensionless), and
depth of soil cover (m).
me?imate of <>* concentration of contanunant in air-fflled
= Ct I [BDKD/H + QJH +; 6J
where:
f uncoveted ** temporarily covered monofill cell are
aYerage rate °f eraissions from a monoful cell during the
-------
monofill (m). Therefore, converting the estimated loss rate (kg/nf-sec) into a first-order loss
coefficient (yr1) requires division by depth and adjustment of units from seconds to years:
"6Xl°7
(7-4)
where:
KV. = first-oider loss rate coefficient for volatilization during facility's
active operation (yr1), and
3.16xl07 = constant to convert units from (sec"1) to (yr1).
Estimated coefficients for losses to volatilization and leaching are combined with assumed
rates of degradation (K,^, obtained for each contaminant from the scientific literature) to yield
a "lumped" coefficient describing contaminant loss through all three pathways during the
facility's active lifetime:
where: ________ ..... , ......
KU = coefficient for total rate of contaminant loss through leaching,
volatilization, and degradation during facility's active operation (yr1).
We calculate the fraction of contaminant loss attributable to each individual process
during the facility's active lifetime as: —
f _ . _ yg _ fg
Jla -if ~ J-va y Jda w-
A*i AM Kta
where:
fu = fraction of total contaminant loss during facility's active operation
attributable to leaching (dimensionless),
fvi = fraction of total contaminant loss during facility's active operation
attributable to volatilization (dimensionless), and
f^ = fraction of total contaminant loss during facility's active operation
attributable to degradation (dimensionless).
The fraction of total loading lost within the facility's active lifetime is calculated
numerically from the lumped rate of contaminant loss, assuming a time step of 1 year:
Mt = 0 (f=0)
Mt =
and:
7-6
-------
where:
4* - fraction of total contaminant lost
where:
"^™nt ior total rate of contaminant loss from inactive monofill
*" = n^fm^^6 °f COntaminant loss ***& volatilization from inactive
The fraction of loss attributable to each individual process is calculated as:
where:
ioss *" jnacave
4 ~ S^H"/ ^ <:OI?aiIlinailt lo» &om faactive monofill attribumble to
degradation (dimensionless).
Methodology for Groundwater Pathway
—
7-7
-------
1) Determine the concentration of contaminant in leachate from the bottom of the
facility, and
2) Use mathematical models for the transport of contaminant through the unsaturated
and saturated soil zones to estimate expected concentrations of contaminant in
groundwater. '
With the mass balance calculations, we have an estimated total rate at which contaminant
is lost from the facility, and the fraction of that loss attributable to leaching. Using methods
discussed in Appendix C, we conservatively estimate the amount of time that would be required
to deplete the entire mass of contaminant deposited in the monofill at the maximum predicted
rate of loss for each contaminant. TMs approach is conservative because using higher estimates
for flux will yield a higher estimate of concentrations at the well.
For monofills, the rate of maximum total contaminant loss (in kg/yr) will occur hi the
year immediately following the last deposit of sludge, since the total mass of contaminant at the
site reaches its peak at that time. As explained in Appendix C, this peak rate of loss could be
maintained for a maximum length of time described by:
TP = LF I [1-e
where:
TP = length of "square wave" in which maximum total loss rate of contaminant
depletes total mass of contaminant applied to site (yr).
This result is combined with the estimate of the fraction of total contaminant loss through
leaching for a conservative estimate of the average flux of contaminant leaching from the
monofill:
10"* C ffa SC
I re
where:
FA, = annual flux of contaminant leaching from the monofill (kg/ha-yr),
C = dry-weight concentration of contaminant in sludge (mg/kg),
SC = estimated mass of sludge contained in one hectare of completed monofill
(kg/ha), and
10"6 = constant for converting units from (mg/ha-yr) to (kg/ha-yr).
Since sludge is combined with soil when disposed in a monofill, the volume (and mass)
of sludge in the monofill is only a fraction of the total volume of the monofill. Therefore, the
dry mass of sludge contained in one hectare of completed monofill is calculated by multiplying
the facility's depth by the fraction of its volume containing pure sludge and by the mass of solids
in one cubic meter of sludge:
7-8
-------
= d^fj MS 104
where:
MS =
and:
SC = estimated mass of sludge contained in 1 ha of completed monofill (kg/ha),
MS = mass of solids in 1 m3 of pure sludge (kg/m3),
f,r = fraction of monofill's volume containing pure sludge (dimensionless)
PA = particle density of sludge (kg/m3),
pw = density of water (kg/m3),
f.oi = fraction of solids in sludge (kg/kg), and
104 = constant for converting units from (kg/m2) to (kg/ha).
Next dividing this estimated flax by the assumed net recharge and adjusting units yields
the estimated average concentration of contaminant in leachate:
0.1 FA,
~~~
0.1 - constant to convert from (kg/ha-m) to (mg/J), and
C^ = average concentration of contaminant in water leaching from the monofill
site (mg//).
The next step is to relate the leachate concentration to the expected concentration of
contaminant in drinking water wells near the site. Two mathematical models are combined to
calculate an expected ratio between these two concentrations. The Vadose Zone Flow and
Transport finite element module (VADOFT) from the RUSTIC model (U.S EPA 1989b c) is
used to estimate flow and transport through the unsaturated zone, and the AT123D analytical
model (Yeh, 1981) is o&ed to estimate contaminant transport through the saturated zone.
VADOFT allows consideration of multiple soil layers, each with homogeneous soil
characteristics. Within the unsaturated zone, the attenuation of organic contaminants is predicted
based on longitudinal dispersion, an estimated retardation coefficient derived from an equilibrium
partition coefficient, and a first-order rate of contaminant degradation. The input requirements
tor the unsaturated zone module include various site-specific and geologic parameters and the
leakage rate from the bottom of the monofill. It is assumed that the flux of contaminant mass
into the top of the unsaturated zone beneath a facility can be! represented by results from the
mass-balance calculations described above. Results from analysis of the unsaturated zone give
the flow velocity and concentration profiles for each contaminant of interest. These velocities
and concentrations are evaluated at the water table, converted to a mass flux and used as input
to the AT123D saturated zone module.
7-9
-------
The flow system in the vertical column is solved with VADOFT, which is based on an
overlapping representation of the unsaturated and saturated zones. The water flux at the
soil/liquid interface is specified for the bottom of the monofill, which defines the top of the
unsaturated zone in the model. In addition, a constant pressure-head boundary condition is
specified for the bottom of the .unsaturated zone beneath tite monofill. This pressure-head is
chosen to be consistent with the expected pressure head at the bottom of the saturated zone,
without consideration of the added flux seeping from sludge in the monofill. Transport in the
unsaturated zone is determined using the Darcy velocity (Vj) and saturation profiles from the
flow simulation. From these, the transport velocity profile can be determined.
Although limited to one-dimensional flow and (transport, the use of a rigorous finite-
element model in the unsaturated 2»ne allows consideration of depth-variant physical and
chemical processes that would influence the mass flux entering the saturated zone. Among the
more important of these processes are advection (which is a function of the Darcy velocity,
saturation and porosity), mass dispersion, adsorption of the leachate onto the solid phase, and
both chemical and biological degradation.
To represent the variably saturated soil column beneath the floor of the monofill, the
model discretizes the column into a finite-element grid consisting of a series of one-dimensional
elements connected at nodal points. Elements can be assigned different properties for the
simulation of flow in a heterogenous system. The model generates the grid from user-defined
zones; the user defines the homogenous properties of each zone, the zone thickness and the
^ number of elements per zone, and the code automatically divides each zone into a series of
"" elements of equal length. The governing equation is approximated using the Galerkin finite
element method and then solved itenitively for the dependent variable (pressure-head) subject
to the chosen initial and boundary conditions. Solution of the series of nonlinear simultaneous
equations generated by the Galerkin scheme is accomplished by either Picard iteration, a
Newton-Raphson algorithm or a modified Newton-Raphson algorithm. Once the finite-element
calculation converges, the model yields estimated values for all the variables at each of the
discrete nodal points. A detailed description of the solution scheme is found in U.S. EPA
(1989b).
One-dimensional advective-dispersive transport is estimated with VADOFT based on the
estimated mass flux of contaminant into the top of the soil column, and a zero concentration
boundary condition at the bottom of the saturated zone. As discussed earlier, sludge is assumed
to be deposited in the monofill for 20 years, followed by an inactive period in which
contaminant is depleted from the monofill by leaching, volatilization, and erosion. To simulate
potential contamination of groundwater, the loading of contaminant into the unsaturated zone
beneath the monofill is "linearized" into a pulse of constant magnitude (TP) to represent the
maximum annual loss of contaminant (in kg/ha-yr) occurring over the 300-year simulation period
modeled. The duration of that pulse is calculated so that contaminant mass is conserved.
As in calculations for the umsaturated zone, degradation of organic contaminants is
assumed to be first-order during transport through the aquifer. Speciation and complexation
reactions are ignored for metals, leading to the possible over- or underestimation of expected
concentrations of metals in groundwater at the location of a receptor well. Detailed descriptions
7-10
-------
°f ^T1230 modd m Provided by U.S. EPA (1986h) and by Yeh (1981) and will not he
Methodology for Vapor Pathway
near tJrito:**8 ***** "" ^^ **" ^ concentration of volatilized contaminants in air
1} c^ta^dn^8 baknCe Calculati?ns summarized above to determine the mass of
acL °f ** L?dUStdal S°Urce ComPlex ^"S Te™ Model
(ISCLT) to model the transport and dispersion of contaminant in ambient air near
me site.
We first use results from mass balance calculations to estimate the fraction of
ontaminant mass expected to volatilize from the monofill within an
where:
fraction of contaminant mass which volatilizes over a human lifetime and
human life expectancy (yr).
mnnnfiH - ti°11 by ** total mais of «>«taminant deposited in the
monofill, and divide by the time of release to calculate an average flux:
LS . . . .,....„
where:
10-* = constant to convert units from (mg/ha-yr) to (kg/ha-yr), and
FA, - annual average flux of volatilized contaminant from the site (kg/ha-yr).
The next step is to relate releases of volatilized contaminant to the exnected
concentrations in amM^n* a;- TT,~ i~i _. ^_ _.. . ^^uuuuu. iu me cxpeciea
monofill site is
7-11
-------
Environmental Science and Engineering (1985). These equations are simplifications of equations
used in ISCLT. The exposed individual is assumed to live at the downwind property boundary
of the monofffl site. A source-receptor ratio is calculated to relate the concentration of
contaminant in ambient air at that individual's location (g/m3) to the rate at which that
contaminant is emitted from the, facility (g/m2-sec):
SRR1. = 2.032
A v
(/•'
u a
where:
SRR
2.032
A
SMA
v
r'
Xy
u
source-receptor ratio (sec/m),
empirical constant,
area of SMA (m2),
sludge management area,
vertical term (dlimensionless),
distance from the SMA center to the receptor (m),
lateral virtual distance to the receptor (m),
wind speed (m/sec), and
standard deviation of the vertical distribution of contaminant concentration
iirambient air (m).
The vertical term (v) is a function of source height, the mixing layer height and
-------
Table 7-1
«
Parameters Used to Calculate az Under Stable Conditions'
x (km)
0.10-0.20
0.21 - 0.70
0.71 - 1.00
1.01 - 2.00
2.01 - 3.00
3.01 - 7.00
7.01 - 15.00
15.01 - 30.00
30.01 -6o;oo
> 60.00
a
15.209
14.457
13.953
13.953
14.823
16.187
17.836
22.651
27.084
34.219
b
0.81558
0.78407
0.68465
0.63227
0.54503
0.46490
0.41507
0.32681
0.27436
0.21716
V Source: Environmental Science and Engineering (1985)
-------
This result is combined with the estimated average flux of contaminant to predict the
average concentration of contaminant in ambient air over this period:
_ FAV SRR
^ 316
where:
C.jr = average concentration of contaminant in ambient air at the receptor
location 0*g/m3), and
316 = constant for converting units from (/*g/m2-sec) to (kg/ha-yr).
7.1.2 Methodology for Surface Impoundment Prototype
Algorithms for Surface Impoundment Prototype
Our methods of estimating'exposure and risk for surface impoundments are similar to
those described in Section 7.1.1 for monofills. As with monofills, we begin with a mass balance
of contaminant losses from the facility.
Methodology for Mass Balance
Contaminants in wastewater sludge are assumed to enter the surface impoundment
through continuous inflow, and to be removed through four general processes:
1) contaminant is lost to degradation within the facility (e.g., to photolysis,
hydrolysis, or microbial decay),
2) contaminant is transported out of the facility by seepage through the floor of the
impoundment,
3) contaminant is lost through outflow (possibly for return to the treatment works),
and
-*) contaminant volatilizes from the liquid surface of the impoundment.
We have adapted our model for describing these four processes from a two-layer model
suggested by Thomann and Mueller (1987) for modeling toxic substances in a lake. For the
water column of a lake, those authors consider the inflow and outflow of contaminant, diffusive
exchange between the sediment layer and the water column, degradation, volatilization, the
settling of paniculate toxicant from the water column to the sediment, and the re-suspension of
paniculate from the sediment layer to the water column. For the sediment layer, they consider
diffusive exchange with the water column, decay processes, paniculate settling from the
overlying water column, re-suspension flux from the sediment to the water column, and loss of
toxicant from the sediment due to net sedimentation or burial.
7-14
-------
time are proportional to the current amcentration of contaminant in the imundme *
We rely on two additional simplifying assumptions:
1) Concentrations of contaminant within each layer are assumed to be at steady-state
and to be partitioned at equilibrium between adsorbed and dissolved pteSs
2) Rates of contaminant transfer and loss when the impoundment is half-filled with
to to ** °f
^
(7-5)
where:
Qi = rate at which sludge enters the impoundment (mVsec)
- concentration of contaminant in inflow to the impoundment (kg/m3)
rate at which outflow leaves the impoundment (mVsec)
"
c-
- rate of contaminanit degradation in liquid layer (sec'1),
A - surface area of imjpoundment (m2),
7-15
-------
di = depth of liquid layer (m),
Kydi = rate of contaminant volatilization from liquid layer (m/sec),
Q«p = rate of seepage beneath the impoundment (m/sec), and
DV = rate of change in the volume of the layer (nrVsee).
Because the total depth of the impoundment (including both liquid and sediment layers) is
assumed constant, the depth of the liquid layer is reduced as more sludge accumulates in the
sediment layer. If the rate at which the sediment accumulates is constant over the active lifetime
of the facility, the rate of accumulation can be determined by dividing the total depth of the
impoundment by its expected active lifetime:
DV = -2—
TF
where:
dti = total depth of impoundment (m),
TF = estimated active lifetime of facility (sec).
The active lifetime of the facility is ejaculated as:
where:
Si = concentration of solids in liquid layer (kg/m3), and
$2 = concentration of solids in sediment layer (kg/m3).
For the first term on the right of Equation 7-5, the volume of outflow from the facility
(Q0) is calculated to be consistent with assumptions about rates of inflow, seepage, and
accumulation of the sediment layer:
The concentration of solids in the liquid and sediment layers is calculated from parameters
describing the percent solids (by mass) in each layer:
5 . ^.-w-i =
1 n D j. ft _ ZIi \ ~ 2
where:
PI = percent solids (by mass) in liquid layer (kg/kg),
P2 = percent solids (by mass) in sediment layer (kg/kg),
pw = density of water (kg/m3), and
P,i = particle density of sludge (kg/m3).
7-16
-------
In each layer contaminant is; partitioned between adsorbed and dissolved phases As
discussed earlier, the partitioning dqjends on both the chemical-specific partition coefficient and
acentration of solids in the laver wenicient and
7 —-— j***—"••"•'• •••"••"fy ***'j.rwu\J
the concentration of solids in the layen
KD Sl
where:
fai — fraction of contaminant dissolved in the liquid layer, and
KD = chemical-specific partition coefficient (m3/kg).
The second term on the right side of Equation 7-5 describes degradation of the
contaminant through photolysis, hydrolysis, microbial decay, and other processes. Values for
fejT,,^? from f"?* °f anMR*ic microbial degradation, and are applied to contaminant
in both di
,,
both dissolved and adsorbed phases.
The third term on the right side of Equation 7-5 describes contaminant loss through
volatilization and is the only term directly linked with human exposure. The overall mis
tiunsfer coefficient for volatilization <*^) is calculated with a two-film resistance model
61' " ^^ ** ^^ resistan<* equak ** sum of
where:
K, = mass transfer coefficient for the liquid layer (m/sec),
Kg = mass transfer coefficient for gas layer (m/sec),
R = ideal gas constant (8.21xlQ-5 atm-nrVKniiol),
T = temperature (K), and
H = Henry's Law constant for contaminant (atm-m3/mol).
Numerous methods for calculating Kt and Kg for water surfaces have been proposed (see
ioof f1? Wang' 1985; MacKay "^ Leiden, 1975; MacKay and Yeun, 1983: Shen
1982; Springer etal, 1984; U.S. EPA, 1987f; U.S. EPA, 1989f). This methodology foUows
an approach described in U.S. EPA (1987f, 19899 for estimating volatilization from surface
impoundments. The selection of appropriate equations for calculating mass transfer coefficients
depends on two characteristics of the site: (1) the ratio of the impoundment's effective diameter
(or fetch ) to its depth and (2) the local average wind speed. Effective diameter (in meters)
is defined as the diameter of a circle with area equal to that of the impoundment. Depth is
defined as that of the liquid layer, which for the purpose of this calculation is assumol to
average half of the impoundment's total depth. The ratio of fetch to depth is therefore calculated
jio« *
7-17
-------
de =
where:
de = effective diameter.(or fetch) of site (m), and
FD = ratio of fetch to depth (dimensionless).
For facilities where the average wind speed 10 m above the liquid surface is greater than 3.25
m/s and FD ^ 51.2 (as in the scenario used for the surface impoundment prototype):
K, = 2.611x10-' «jj IDJDJP (7-7)
where:
u10 = average wind speed 10 m above surface (m/sec),
Dcw = diffusivity of contaminant in water (cm2/sec), and
D^ = diffusivity of diethyl ether in water, 8.5 x-10* cmVsec.
Calculation of the mass transfer coefficient for the gas phase is based on Hwang (1982).
For all values of FD and U10, Kg (m/sec) is calculated from
Kg ;-= l.SxlO-3^78 Sc?*7 de^n
where ScG equals the Schmidt number on the gas side, defined as
and where:
M. = viscosity of air (g/cm-s),
P. = density of air (g/cm3) and
D,., = diffusivity of constituent in air (cm2/sec).
Equations 7-6 through 7-8 are sufficient to estimate K^, the overall mass transfer coefficient
for the dissolved fraction of the contaminant.
The fourth term on the right side of Equation 7-5 describes losses of dissolved
contaminant from the liquid layer as a result of the seepage through the sediment layer and the
floor of the impoundment. The rate of seepage (Qse^ is based on measured values from sludge
lagoons. Only dissolved contaminant is included hi this term; adsorbed contaminant is included
in the fifth term of the equation, which describes loss of contaminant from the liquid layer as
a result of the diminishing volume of that layer.
7-18
-------
so that:
DV
£»ntaminant lost from liquid layer that is lost in outflow
^ trom the impoundment (dimensionless),
W - fraction of total contaminant lost from liquid layer that is lost to
degradation (dunensionless),
fvoa = fraction of total contaminant lost from liquid layer that is lost to
volatilization (diaaensionless),
f«p, = fraction of total ccntaminant lost from liquid layer that is lost to seepage
(dimensionless), suid »"*«&*
fddi = faction of total contaminant lost from the liquid layer as a result of the
diminishing volume of the liquid layer.
tM. i Se^ment Lay"F- Contaminant mass accumulates in the sediment layer as the depth of
this layer increases and eventually reaches the surface of ttae impoundment If the oMyTource
of contaminant mass for the sediment layer is the loss estimated for the liquid layer:
C, * DVC2
where:
fd2 = fraction of total contaminant in sediment layer that is dissolved
(dunensionless) ,
d2 = depth of sediment layer (m),
rate of contaminant degradation in sediment layer (sec'1), and
total concentration of contaminant in sediment layer (kg/'m3).
7-19
-------
A coefficient for the total loss or storage of contaminant in the sediment layer (£^2, in m3/sec)
can be defined as:
As with the liquid layer, this coefficient can be partitioned into its individual components:
f = eg f sept DV
**2 " K
where:
fdeg2 = fraction of contaminant reaching the sediment layer that is lost to
degradation (dimensionless),
f«p2 = fraction of contaminant reaching the sediment layer that is lost to seepage
(dimensionless), and
fdeE = fraction of contaminant reaching the sediment layer that is stored in the
accumulating depth of this layer (dimensionless).
If concentrations of contaminant in the liquid and sediment layers can be approximated as steady-
state for the duration of the impoundment's active lifetime, and if the partitioning of contaminant
among competing loss processes halfway through the impoundment's active lifetime is assumed
typical of its entire active phase, then the fraction of each year's loading of contaminant lost
during each year of the facility's active phase (f^ can be calculated as:
fact ~ fvoll * fdegl + foutl +
Finally, if all contaminant is eventually lost from the impoundment and the partitioning of
contaminant mass halfway through the facility's lifetime is generalized for the entire mass of
contaminant, the fraction of contaminant mass lost through each pathway can be calculated
as:
7-20
-------
+J
-------
This result is combined with another result from the mass balance calculations to derive a
conservative estimate of the average flux of contaminant to the unsaturated zone beneath the site:
c
TP
»
where:
0.01 = constant to convert units from (mg/m2-yr) to (kg/ha-yr),
FA, = average flux of contaminant seeping through the floor of the surface
impoundment (kg/ha-yr), and
C = concentration of contaminant in sludge (mg/kg).
Next we use the average flux to estimate the average concentration of contaminant in
seepage:
C
3.2x10-'
-------
contaminants through the saturated zone, with adjustments to represent increased elevation of
e wa er Die. Unlike RUSTIC, however, the present methodology does not allow for nartial
feedback between the unsaturated and saturated zone components of the modeffcesatS
zone is represented separately by an analytical transport model. saturated
Saturated Zone
f * ^ Ai123D m0del aCCeptS 3S mput ±& flux of Pure contaminant mass entering the top
f^^if4^ Z°ne' *"? ^ n0t <:onsider ^ extent of the contaminant's dilution by water
from the source area, or the impact of that water on groundwater flow within the saturated zone
When the vertical movement of contaminant through the unsaturated zone is due
infiltration throughout the area, the gradient within the aquifer is a function of the water ent,
the saturated zone, and neglect of the .diluted state of the source term may be valid. For the
of a surface impoundment, however, neglect of the extent of the contaminant's original dil,
could result in non-trivial overestimation of the source concentration leading to an
^SSSSL?^^tam^t concentrations at the receptor weU. Furthermore, neglect of
the velocity of
^
Df - Fc i<4Q^+FJ (7-9)
where:
F. = ^ volume of fluid passing through a vertical cross section of the aquifer
oriented perpendicular to the direction of flow, and having a width equal
to the source width and a depth equal to the saturated thickness of the
aquifer (m3/sec).
The excess water released by seepage from a surface impoundment can also result in a
superimposed radial velocity field on the background or regionT velocity field of^^er
S f°*er,WOldS' the, h0riZ°ntal VdOCity Of water within *» a^er can bellowed \£
gtadie... of the lagoon, and accelerated downgradient of the site. Thic change in the velocity
field might result in reduced time of travel for contaminants moving to receptor w2
£r?^ient ° ^Po^^ent site, which could in turn lead to reductions iTSntaminant
SS£?°n Pn°L hlfinanfie?)osure- Accur*e accounting; of the influence of mixing and
degradation would require a fully three-dimensional flow and transport model; this methodoloffv
uses a simpler approach to estimate a conservative limit to contaminant decay within the system
The limit is estimated by increasing the estimated velocity of groundwater flow to account for
toe maximum downgradient increase in velocity due to the source. The velocity increase can
be approximated by idealizing the lagoon as a circular source, so that the rate at which seepage
passes outward through a cylinder ben^th the perimeter of the lagoon's floor is-
7-23
-------
where:
V; = superimposed radial velocity from water seeping from impoundment
(m/s), and
d, = depth of aquifer (m). •
In addition to increasing the expected velocity of contaminant transport through the
aquifer, this superimposed velocity would also have the effect of increasing AT123D's estimate
of contaminant dilution within the aquifer. This additional dilution effect must be subtracted
back out of the model calculations, since the true dilution is explicitly included in the factor
introduced by Equation 7-9. The model performs this calculation automatically, based on the
following equation for the anti-dilution factor:
where:
D,f = anti-dilution factor,
vv = the vertical velocity due to the source (m/s), and
vh - = the regional velocity of horizontal groundwater flow (m/s).
It should be noted that the above methodology is conservative, since it overestimates the
velocity beneath the source and dcxjs not allow for decreases in the superimposed velocity
beyond the source. As a result, the methodology is more conservative than a three-dimensional
model. In comparison with a two-dimensional cross-sectional flow and transport model, the
model is more conservative beneath the source, but less conservative beyond the source.
By combining the VADOFT model with AT123D, and by adjusting calculations in
AT123D to accommodate the dilution and superimposed velocity described above, concentrations
of a contaminant in groundwater at a receptor well can be predicted as a function of the liquid
concentration of contaminant near this floor of the impoundment, the rate of seepage from the
facility, and hydrogeological characteristics of the area. It should be noted that all of the
calculations described above are tinear with respect to contaminant concentrations in liquid
seeping from the lagoon.
This result is combined with chemical-specific properties and assumptions about the
physical characteristics of the unsaturated and saturated zones to provide inputs for the linked
VADOFT and AT123D models. Model results include an estimate of the maximum
concentration of contaminant (C^) expected in a well at the downgradient edge of the site's
property boundary within the 300-year period simulated.
7-24
-------
Methodology for Vapor Pathway
""
tssszs=*zssxss&:£«~
^ ^vin' 77 / f
»J«^^^.iV/ *^a j /
• . . ^ _ 7^ tftff *voll
TF
where:
h = ^c?onoftota][ contaminant volatilizing during human lifetime
^ ^ = hfetune expectaincy (yr), and '
3.2xl07 = constant to convert units from (yr) to (sec).
site as: ™8
°°nVerte
-------
of sludge surface disposed, but lacks information describing characteristics of individual disposal
sites. For simplicity and for kck of site-specific data, we have chosen "reasonable worst case"
scenarios to represent the monofill Jind surface impoundment prototypes. Only two types of
parameter values are assumed to vaiy from site to site: concentrations of contaminants in the
sludge and the density of human populations surrounding the site.
Estimated concentrations in environmental media are converted to estimates of human
exposure based on assumptions about the rate at which the average individual consumes drinking
water and inhales air. We calculate potential exposure through ingestion of contaminated
groundwater as:
EXP, = -sfi
J BW
where:
EXPj = human exposure to pollutant./ (mg/kg-day),
BW = average body weight (kg),
Iw = quantity of water ingested daily (J/d), and
Cwel = concentration of contaminant in well water (mg/l).
For air, we calculate human exposure as:
BW
where:
10"3 = constant to convert units from 0*g) to (mg),
G.J, = average concentration of contaminant./ in air at the receptor site 0*g/m3),
and
I. = inhalation volume (nrVday).
To complete our calculations; of risk, we combine estimates of individual exposure
through groundwater and ambient air with estimates for the sizes of exposed populations.
Finally, we scale by the estimated number of surface disposal sites in the U.S. to calculate our
estimates of total risk.
7.2 DATA SOURCES AND MODEL INPUTS
As described in Section 7.0, we base our estimates of aggregate risks on separate
prototypes for surface disposal: a monofill and a surface impoundment. With data from the
analytic survey of the NSSS, we model 26 sites to represent an estimated 1,446 surface disposal
facilities. Of these, 12 are modeled as monofills (including 3 dedicated land-application sites
which were re-classified as surface disposal) and 14 are modeled as surface impoundments.
Because reliable data could not be obtained for key site-specific parameters at many of the sites,
7-26
-------
for
7.2.1 Site and Sludge Parameters
Area of Surface Disposal Site
.Depth of Disposal FaciHty
,th f * A?0™*^ d^eimines tototal q^tity of sludge contained in the site The
•in or j.40 m is based on a design scenario described in U.S. EPA (l9S6h) The
the largest number of surface impoundments in the 1988 W>Ssurve^
Distance to Well
that oneF°r ^tor^iTuS "^ ^^ !mp?undment Prot<*ypes, we conservatively assume
the site 050 m ^^^^^^^l!!!^^.^1^^^!^ ^own-gradient of
ited for this location are
the site, and within a 90
S*»Jfifi£SS£S&^.^
1000-3000 m.
Thickness of Cover for Monofill
nf ffi<» fin t 1 Tnu «~«.w»«*i. *j aoouui&u iu uo v. j m, auQ UlC U
or me rinai cover 1 m. These values represent typical thicknesses for cover applied to
7-27
t0
an area
-------
Table 7-2
Site 'and Sludge Parameters
Monofill Prototype for Surface Disposal
*
Parameter ' Value
Area of monofill (m2) , 10,000*
Depth of monofill (m) 3.46'
Distance to Well (m) 150>>
Thickness of Daily Cover (m) 0.3b
Thickness of Permanent Cover (m) lb
Time Average Cell Contains Sludge (hr) 87,660°
Sludge as Fraction of Total Volume (nrVm3) 0.63b
Active Site Life (yr) 20b
Wind Velocity (m/sec) 4.5d
Average Air Temperature (K) 288*
Solids Content of Sludge (kg/kg) Q.2&
* U.S. EPA (1986h).
b U.S. EPA (1978).
c Half of assumed site life.
d U.S. EPA (1990c).
7-28
-------
Table 7-3
•t Site and Sludge Parameters
Surface Impoundment Prototype for Surface Disposal
value
Area of Surface Impoundment (m2) 20 236*
Depth of Surface Impoundment (m) . 4b
Distance to Well (m)
Rate of Inflow (m3/sec) n ,
i u>v
Wind Velocity (m/sec)
Average Air Temperature (K) 288
Solids Content of Inflow (kg/kg) 0 03d
SoUds Content of "Liquid" Layer (kg/kg) 0.03d
Solids Content of "Sediment" Layer (kg/kg) 0.175c>
Particle Density of Sludge (kg/m3) 1200*
1 U.S. EPA (1986f). ~~
b Abt Associates Inc. (1989).
c U.S. EPA (1990c).
d U.S. EPA (1978).
e Chaney (1992).
7-29
-------
fill trench (U.S. EPA, 1978).
Number of Days Average Cell of Menofill is Uncovered
Total emissions from a monofill cell depend on the length of time the eel is uncovered
or covered with a soil layer. For these calculations, we assume a cell is open for 4 hours a day
for the 3 consecutive days it receives sludge (U.S. EPA, 1986h). For the remaining 20 hours
of each of those 3 days, and for the remainder of the active lifetime of the facility, the cell is
covered with a temporary layer of soil. Assuming a constant rate of disposal, during the 20 year
active lifetime of the monofill, half the cells will contain sludge for more than 10 years and half
for less than 10 years. A typical cell is therefore uncovered for (3x4)/24 or 0.5 days and
covered for (20x365.25)72 - 0.5 or about 3652 days. After the facility is filled to capacity, it
is covered with a thicker, permanent layer of soil for the remainder of the period simulated.
Inflow Rate for Surface Impoundment
We assume the prototype faciility for surface impoundments receives continuous inflow
of sludge throughout its active lifetime. The duration of that active lifetime depends on the rate
at which sludge enters the facility, the solids content of that sludge, and the volume of the
facility. According to the RCRA Subtitle D survey (U.S. EPA, 1986f), almost 96 percent of
the sludge lagoons surveyed received less than 50,000 gallons per day of wastewater flow. This
number converts to 0.0022 mVsec, and is used as the model inflow rate. It is also consistent
with the mean inflow rate for surface impoundments, based on data from the analytic survey of
theNSSS. -----
Ratio of Sludge to Total Volume
A typical trench monofill contains parallel trenches filled with sludge and separated by
soil, so that the entire monofill site contains both sludge and soil. We do not model these layers
of sludge and soil separately, but instead model an idealized, homogeneous mixture of sludge
and soil. We calculate the total quantity of sludge likely to be contained within a monofill of
specified dimensions from the fraction of the monofill's contents consisting of sludge. U.S.
EPA (1978) describes several design scenarios for different types of trench monofills, and
reports the approximate quantity of slludge that can be received per acre for each scenario. The
wide trench monofill scenario (which receives the most sludge per unit area) is reported to
receive about 7,744 m3 of sludge per hectare to a depth of about 1.7.2 m. Dividing this volume
of sludge by the volume of facility per hectare (10,000 m2 x 1.22 m or 12,200 m3) yields the
fraction of the monofill's volume wMch contains pure sludge (0.63).
For the surface impoundment prototype, we assume the entire facility is filled with
sludge, so that the volume of sludge contained by the facility is equal to the volume of the
impoundment.
7-30
-------
«fe Life
«• a» £,*?*
design
£*B.*«S17i*
^ ^p^Tw *" Ifci^l^«^a3* ssiS'^S^'
m"a**«w^
1716 solids con* ^ *lde tmnlt ^ "^ttmerf fn!:? a^ent fe / ^P0^
zed IOT,^ Content nf «i "coco mnnrt^-i. r°rt&e m ' •^-, so/h
^* Dyers' n "i- Of sludoa • "uonofiU fn c J^ Qionafin *v
'ftd *_ " ™ -UOllirf" »_ SC in c.._r. VJ-O. Rt>A ._•**« Dnn*«^
;g>-t^S^i-
ac* Bionofiij m c^J?e «»oaofin ' *%> to
( >S-^A, I97[f Pt>fotj7>e
^^^umerf.,...
„ . n
-------
Particle Density of Sludge
A particle density can be derived from the mass and solids content of a typical wet
sludge. According to Chaney (1992)', a typical wet sludge in a surface impoundment has a
specific gravity of about 1.03, equivalent to a density of about 1030 kg/m3. If 17.5 percent of
the mass of such sludge is solids'and the remainder is water (with a density of 1000 kg/m3), the
particle density of pure sludge (p^ ican be calculated as:
Concentrations of Pollutants in Sludge
Table 7-4 lists the concentrations of each pollutant measured for each of the 26 surface
disposal facilities included in the analytic survey of the NSSS and modeled for this analysis.
7.2.2 Soil and Hydrologic Parameters
The unsaturated zone is characterized by pore space containing both air and water,
whereas the pore space hi the saturated zone contains water only. Because of differences in fluid
flow regimes, these two zones require different equations and input parameters for tracking
contaminant transport. A simplifying assumption used for this analysis is that the basic soil
characteristics (including soil type, irorosity, and bulk density) of the two zones are identical.
For both the monofill and surface impoundment prototypes, sludge is assumed to be
placed entirely beneath the ground surface, requiring subsurface excavation. For monofills, the
excavated soil is typically used for cover (U.S. EPA, 1978). Parameter values describing soil
and hydrogeological characteristics for both monofills and surface impoundments are listed in
Table 7-5.
Soil Type
The types of soil in the unsaturated and saturated, zones affect the ability of a contaminant
to move vertically to the aquifer atnd laterally to a nearby well. In general, the ease of
contaminant transport through a soil (ignoring the adsorption properties of the soil) is largely
affected by the type of clay present, the shrink/swell potential of that clay, and the grain size
of the soil. The less the clay shrinks and swells and the smaller the grain size of the soil, the
-------
1
c
E
0
j
1 2
. I 1
a
' 1 1
• I 3
GO
Q
"f § 1
*~- _ »
<='«^^0^0iSo^^
* - d vo ^
^
ft,
§3 • a
P* Ja 8 r °
to.* ^ t S '
sllo^l^^l^SsSf
I i i i i ii,r*fj i *
2<3£O eg- s; £ rS-^jrS ^^S 5 S
- * vi -z a* c&. s # a g jg
-------
1
o
w
•t
J
9)
Is
P*M
t
to
8
•
'o
rt
o
ex
ca
3
i
*. U>4
03
'«5
*••
1
S
1
*©
V]
0
*^*
4— *
fi
"*-»
1
C
O
U
cu
i
8
1
1
2
>>
I
i
1
fe
Q*
a
£
'i
n
6
"3
•g
.2
5
52
Q
£
00
8
d
8
d
8
d
o
o
o
VO
I—I
eN
«s
O
' O
oo
H
1 |
Si |
< D &
|2 1
oo ^ 55
8
O
1—4
8
d
*•*
O>
t-
""'
ts
3
^
o\
pj
00
^
s
cu
.c
e
'I
8
d
o
8
d
en
en
VO
oo
H
VO
JO
1
10
vd
CM
oo
£
£
c/5
1
>
0
1
8
d
8
• d
8
d
o
0
en
en
cs
o
""
en
^
O
fc
^
^
5
1
tn
s
8
O
^
2 '
8
d
o
00
d
C~;
tn
en
vq
en
«o
en
en
en
c-i
^
O
1|
aS
8
d
i
8
d
JO
"
>o
en
S
00
r--
00
en
os
vo
£
fa
r-!
Idenville Soutl
o
8
d
00
8
d
vo
•o
s
-
««
s
en
*o
Tf
VO
(^
TJ.'
OS
(S
^
{J
%
ii
8888
o o d d
t •* es o o
^- oo eM Q
en >o 5i
c^
to « «o Os
*°. <* vo S
os en
vd °°. . es °i
l™t
a. *OT
(^ o
oo £
1 •* •==
1 & 3 o. 1
r" OO t. 3 _
2 | w g
"^ c 55 *-< «c
33 • •< w ^ O*
I
d
it
es
d
-------
Table 7-5
, Soil and Hydrologic Parameters
For Monofill ,and Surface Impoundment Prototypes
Parameter
Soil Type
Porosity of Sludge/Soil
Porosity of Soil Cover (monofill only)
Bulk Density for Pure Soil (kg/m3)
Bulk Density of Sludge/Soil (kg/m3)
Saturated Hydraulic Conductivity of Soil (m/hr)
Water Retention Parameters
9,
a (m-1)
ft
Fraction of Organic Carbon in Soil or Sludee
Sludge
Unsaturated Soil Zone
Saturated Soil Zone
Depth to Groundwater (m)
Net Recharge or Seepage
Monofill Prototype (m/yr)
Surface Impoundment Prototype (m/yr)
Thickness of Aquifer (m)
Hydraulic Gradient
Value
Sand
0.4«-b
0.4**
1600°
1400*
0.61b
0.045*
14.5b
2.68"
0.316
0.001f
0.5f
2.5«
5
0.005f
' Todd (1980).
b Carsel and Parrish (1988).
a979)Ulated from porosity "^particl
-------
model simulations are based on values estimated for sand.
Porosity of Sludge/Soil
Porosity is the ratio of ttje void volume of a given soil or rock mass to the total volume
of that mass. If the total volume is represented by Vt and the volume of the voids by Vv, the
porosity can be defined as 0t=Vv/Vt. Porosity is usually reported as a decimal fraction or
percentage, and ranges from 0 (no pore space) to 1 (no solids).
For this analysis, we assume a total porosity of 0.4, based on Todd (1980). This value
is consistent with the average value for sand (0.43) reported in Carsel and Parrish (1988). It
is used to represent total porosity within a monofill, in the cover soil applied to a monofill, and
within the unsaturated and saturated soil zones beneath both monofills and surface impoundment
prototypes. In the unsaturated zone, about half of total porosity on average is assumed to be
water-filled, so that both water-filled and air-filled porosities ares assigned values of 0.2.
Effective porosity is calculate as the difference between the average saturated water
content and the approximate average residual water content, and refers to the amount of
interconnected pore space available for fluid flow. For these calculations, the average residual
water content in the unsaturated zone is assumed to be less than 0.05 (Carsel and Parrish, 1988),
and effective porosity has been approximated with the same value used for total porosity (0.4)
in mass balance and groundwater transport calculations.
Bulk Density of Soil
The bulk density of soil is defined as the mass of dry soil divided by its total (or bulk)
volume. Bulk density directly influences the retardation of solutes and is related to soil
structure. In general, as soils become more compact, their bulk density increases. Bulk density
can be related to the particle density and porosity of a given soil as:
BD = pM(l - Q)
where:
BD = bulk density of soil (kg/m3),
PK = particle density of soil (kg/m3) and
6t = porosity of soil (dimensionless).
Typical mineral soils have pairticle densities of about 2650 kg/m3 (Freeze and Cherry,
1979). This value and a soil porosity of 0.4 suggest a bulk density of about 1600 kg/m3 for pure
soil, somewhat higher than the 1300-1500 kg/m3 range typically encountered for soil mixed with
sludge (Chancy, 1992).
7-36
-------
Saturated Hydraulic Conductivity of Soil
Saturated hydraulic conductivity refers to the ability of soil to transmit water which is
governed by the amount and interconnection of void spaces in the saturated^*
may occur as a consequence* of inter-granular porosity, fracturing, beddi
macropores. In general, high hydraulic conductivities are associated with r
contenunant transport. We use a value for saturated hydraulic conductivity (0.61
on tne yjw percentile of a nmhahilitv rfictntnifinn A**1»,^_..» i.._^. •/ .
on/i DO*-*: h /moo\ %«.. i- «v»»«»*u*v wuuuv,uvu.y m sanu aenveu o\
ana pamsh (1988). This value thus represents a conservative or "reasonable worst case"
Unsaturated Hydraulic Conductivity of Soil
*" hydraulic 'O'Kfoctivity, which is based on the effective
torf Properties, is a function of the moisture content, which is in mm
thS T rSS!!f head ™ese relations^ are central to the simulation of water flow
tough the unsaturated zone. As input, the VADOFT model accepts sets of date poS
descnbmg effective penneability-saturation curves and the saturati^L^ure LS cu^f
%5FS!£'2 **?** VEn GenUChlen water "to"*" Parameters defining the cSves ml'
EPA, 1989b; Carsel and Parrish, 1988); this latter option is used for this analysis
utor r^ fl°m the Soil Con^nration SuTvey (SCS), Carsel and Parrish derived
fir^- ?r *» P"««*» required (0r, «, and ^according to twelve SCS textural
^fd 2 ^ for«( ^ PmMl'. 1988)" Values used for our calculations (0.045, 14.5 m^
and 2.68 for 6r, a, and 0, respectively) correspond to values reported for sand.
Fraction of Organic Carbon in SoU or Sludge
nro™^^^1^*™5*?* fraction Of OIganic carbon * ^ soil with each contaminant's
organic carbon partition coefficient to determine the partitioning of contaminant between soil and
Ton^n ,gen<5?' a 10*6r ^^ °f OIganic carbon ^Plies greater Ability for organic
v^r^ • OIgaU!f C3lb0n ***** for sludSe varies ^ong sludge types, with mean
values for vanous types showing a relatively narrow range of 27.^32.6 percent (U.S. EPA
^!K; J0r*1ld8e Wlthin ^ SUlface disP°saI facility, we use the mean value for all sludges
combined of 31 percent (U.S. EPA, 1978). We have selected a value of lO* for the naction
of organic carbon in the unsaturated zone because it is a typical value for sand and iS a"
faciMesm11? ^ S/0™1 J 0-?1} "f*** f°r !SOil bBneatl1 hazardous waste disposal
raciuties (U.S. EPA, 1986f). The fraction of organic carbon in the saturated zone is expected
to be lower than that of the unsaturated zone, and has been assigned a value of 104, or one-tenth
the traction assumed for the unsaturated zone.
Depth to Groundwater
The depth to groundwater is defined as the distance from the lowest point of the surface
djosal facility to the water table. - The water table is itself defined as the subsurface boundary
between the unsaturated zone (where the pore spaces contain both water and air) and the
7-37
-------
saturated zone (where the pore spaces contain water only). It may be present in any type of
medium and may be either permanent or seasonal. The depth to groundwater determines the
distance a contaminant must travel Ibefore reaching the aquifer, and affects the attenuation of
contaminant concentration during vertical transport. As this depth increases, attenuation also
tends to increase, thus reducing, potential pollution of the groundwater.
For both monofills and surfare impoundments, we conservatively assume that the depth
to groundwater is 1 m. According to data in the GRNDWAT data base (U.S. EPA, 1988a), this
value is less than any of the depths most typical of counties containing surface disposal facilities
in the analytic survey of the NSSS. For this reason, 1 m is believed to be a reasonable worst-
case value likely to over-estimate actual exposure and risk.
Net Recharge or Seepage
The primary source of most groundwater is precipitation, which passes through the
ground surface and percolates to the water table. Net recharge is the volume of water reaching
the water table per unit of land, and determines the quantity of water available for transporting
contaminants vertically to the water table and laterally within the aquifer. The greater the
recharge rate, the greater the potential for contaminant transport, up to the point at which the
amount of recharge is large enough to dilute the contaminant. Beyond that point, the effect of
the increased rate of transport is offset by dilution (U.S. EPA, 1985b).
For monofills, the selected recharge rate (0.5 m/yr) represents the average of a range of
values presented in (U.S. EPA, 1986f). For surface impoundments, the relatively high water
content of sludge can provide an additional source of recharge if water from the sludge seeps
through the floor of the impoundment. In impoundments receiving continuous or periodic
deposits of sludge, this source may not be depleted during the active lifetime of the facility.
Table 7-6 lists seepage rates from municipal lagoons (in inches per day and l/m2-hr) (U.S. EPA,
1987e). The value selected for this analysis (2.5 m/yr) represents the average seepage rate for
lagoons over sandy soil,
Thickness of Aquifer
Saturated zones are considered to be aquifers if they can transmit significant volumes of
water. Only aquifers are considered when selecting input parameters for these calculations. For
estimating aggregate risks, we assume the thickness of the aquifer is 5 m.
Hydraulic Gradient
The hydraulic gradient is a function of the local geology, groundwater recharge volumes
and locations, and the influence of withdrawals (e.g., well .fields). It is also very likely to be
indirectly related to properties of porous media. Rarely are steep gradients associated with very
high conductivities. No functional relationship exists, however, to express this relationship.
The hydraulic gradient value selected for our calculations is 0.005 m/m or 0.5 percent,
and is based on an average value for groundwaters surveyed for the Hazardous Waste
7-38
-------
Table 7-6
Summary of Measured Seepage Rate from Municipal Lagoon Systems1 .1
Water Depth
(ft)
5
6
5
6
6
-
-
-
-
5
5
5
* Qrtiiw»*»« TT C LJ1
Lagoon Type
Facultative
Facultative
Facultative
Facultative
Facultative
_
Maturation
Facultative
Facultative6
Evaporation*1
Facultative
DA /"JrioT^X
Underlying Soil
Heavy silty clay
Light silty clay
Alkaline silt
Fine sand
Gravel and silt
Sandy soil
Sand and gravel
Sandy soil
Clay loam and shale
Mica and schist
Silt, sand, marl
Sand, silt, marl
Sand, silt, marl
Sandy soil
Seepage Rate
(in/day)
0.3
0.29
0.65
1.2
1.3
0.35
0.61"
0.34
0.3
0.06 - 0.23
0-18
1.07
0.04-0.11
0.12
Seepage Rate
(//m2-hour)
0.32
0.31
0.69
1.3
1.4
0.37
0.65
0.36
0.32
0.06 - 0.24
0.19
1.13
0.04 - 0.12
0.13
b Includes net precipitation/evaporation.
6 Used intermittently.
d Sealed with bentonite and soda ash.
7-39
-------
Management System Land Disposal Restrictions Regulation (U.S. EPA, 1986f). .
7.2.3 Chemical-Specific Parameters
• ' •
Distribution Coefficients *
Contaminant transport in soil systems is influenced by interactions between the
contaminant and soil. The affinity of contaminants for soil particles may result from ion
exchange hi clay particles, electrostatic forces between contaminants and charged particles, and
interactions with organic carbon. When all interaction and exchange sites in a soil are filled,
soluble contaminants will move through the soil at the same velocity as the bulk leachate. The
affinity between a soil and a contaminant is characterized by the distribution coefficient (KD).
Representative KD values (in //kg or m3/kg) are defined as the equilibrium ratio of the
contaminant concentration hi soil (mg/kg) to that in associated water (mg/l or mg/m3). Values
used for this analysis are listed in Table 7-7 and discussed below. Note that the organic
contaminants nitrosodimethylamine and toxaphene are not being considered in this analysis
because, they were never detected in the analytic survey of the NSSS.
For hydrophobic organic contaminants, KD is calculated from a contaminant's partition
coefficient between organic carbon and water:
KD
where:
KD = equilibrium partition coefficient for contaminant (m3/kg),
KOC = organic carbon partition coefficient (m3/kg), and
fw = organic carbon as a fraction of soil mass (dimensionless).
As discussed previously,^,, values of 0.31, 0.001 and 0.0001 are assumed for the sludge layer,
unsaturated zone, and saturated zone, respectively.
The organic carbon partition coefficient for a contaminant can be estimated from its
octanol-water partition coefficient, which can be measured in laboratory experiments. Values
of KOC used are shown in Table 7-8, and are calculated from the following regression equation
by Hassett et al. (1983):
log(KOQ = 0.0884 +0.909 log(KOW)
where:
KOW = octanol-water partition coefficient for contaminant.
With the exception of PCBs, the KOW values used for this analysis have been obtained from the
CHEMEST procedures in the Graphical Exposure Modelling Systems (GEMS and PCGEMS),
U.S. EPA (1988a, 1989d).
7-40
-------
Table 7-7
'. *
Arsenic
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Chlordane
DDT
Lindane
Polychlorinated biphenyls
Trichloroethylene
Within
Surface Disposal
Facility
0/kg)
20
431
59
98
621
330
63
32.8
139,000
16,800
41,200
239,000
726
467,000
60.1
Unsaturated
Zone
(ifcs)
20
431
59
98
621
330
63
0.106
448
54.1
133
772
2.34
1,510
0.194
Saturated
Zone
tf/kg)
20
431
59
98
621
330
63
0.0106
44.8
5.41
13.3
77.2
0.234
151
0.0194
Notes: The distribution coefficient for organic pollutants (KD) is the product of the organic
carbon partition coefficient (KOQ and the fraction of organic carbon in the medium (£.).
Assumes fx of 31 percent within the surface disposal facility and 0.1 percent and OJ01
percent in the unsaturated and saturated soil zones, respectively. Distribution coefficients
for metals are geometric means of values reported for a "sandy loam1' soil in Gerritse et
al (1982).
7-41
-------
Table 7-8
Octanol-Water and Organic Carbon Partition Coefficients
*, for Organic Contaminants
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Chlordane
DDT
Lindane
Polychlorinated biphenyls"
Trichloroethylene
Log of Octanol-
Water Partition
Coefficient*
2.13
6.12
5.11
5.54
6.38
3.61
6.70
2.42
Organic Carbon
Partition
Coefficient6
106 .
448,000
54,100
133,000
772,000
2,340
1,510,000
194
* All values except for PCBs taken from the CHEMEST procedure of the Graphical
Exposure Modeling System (GEMS), U.S. EPA (1989d).
b KOC for organic contaminants derived from KOW with Equation 6 from Chapter 15 of
Hasset et al. (1983): log(£OQ = 0.0884 + 0.9091og(JK?W).
c Based on Aroclor 1254, the most common PCB mixture in sewage sludge. Derived from
O'Connor (1992) and representative values from Anderson and Parker (1990).
7-42
-------
Polyctdonnated biphenyls (PCBs) are a class of chemicals containing 209 possible
congeners. The most common POT mixture is Aroclor 1254, which is dominated brpenta-
congeners, with about equal amounts of tetra- and hexa-congeners. In a well-aeed soil
±ZS" S W;*PCBs' l"1?0?*' A**101" 12<®> which contains more penta- and hexa-
congeners than tetra-congeners, is more representative of the PCBs found (O'Connor 1992)
because theless chlorinated congeners degrade more rapidly. To determine a representative
organic carbon partition coefficient for PCBs, we have calculated an average from log KOW
coefficients hste^ in Table 7-9 (from Anderson and Paricer, 1990). The log KOW for the penta-
congener has been estimated to be approximately 6.5 by noting that the loe KOW values are
approximately linearly related to the number of chlorines in the congener. Averaging that value
with the hexa-congener value gives 6.7 for the log KOW. As with other organic contaminants
the regression equation from Hassett et a/., (1983) is used to convert this KOW value to
estimate
For metals, values for KD are taken from Gerritse et al. (1982), and represent results of
laboratory column tests with a sludge-amended sandy loam topsoil. The values used for each
J ta TaWe 7'6' - " ^ °n *» •—< — «* «* ranges
Degradation
concentfat;ons * the subsurface regime may be decreased by various
Pf£?SSel' ****?* abi0ti° hydrolysis ** ***** or Aerobic microbial
^f°Ugh ^f of hydrolysis •» dePen^t only on Ph and temperature (and can
with reasonable accuracy), estimates of rates for microbial degradation are fraught
-r^/ un?&Itafty is due to mafly confounding influences in the field, such as
ilability (fraction of organic carbon present), temperature, the microbial cohsortium
and microbial acchmation to a given pollutant. Nevertheless, the range of microbial degradation
mtes obtained in the laboratory by measuring the rate of disappearance of a pollutant if various
?^^ ' s°acol'amnst"dies, etc., provides a rough estimate of the rate that
activity is hkely to degrade a particular poUutant in the field.
HP^H ^ ShTD ^U?"8 7~10' ^ W0ric Utili2es "VBnd *?««» for representative microbial
degradation rates Where a range of values is reported by these sources, values from the lower
endoftiie range have been selected to derive estimates most protective of public health. Studies
of biodegradahon in soil have been favored over studies of biodegradation in aquatic
Snt^ff ' i, t^?68 °f Only aei°bic W°*gndatfan rates are avaikble for a given
contaminant, a half-life for anaerobic biodegradation has been conservatively estimated to be
four times onger (Howard* a/., 1991). However, if avaikble data fail to show any indication
regime' a value of ° has been assumed for the microbial
am taJ^^^8^
-------
Table 7-9
• «
Octanol-Water Partition Coefficients
forPCBs*
Congener
2,4'
2,2',5,5'
AUPenta
2,2',4,4',5,5'
Average"
Number of Chlorines
2
4
"5,
6
5.5
LogKOW
5.1
6.1
6.5b
6.9
6.7
* Source: Anderson and Parker (1990).
b Estimated based on apparent linear relationship between number of chlorines on congener
and log KOW.
0 log KOW values for penta- and hexa-congeners averaged for representative log KOW.
7-44
-------
Table 7-10
Degradation Rates
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)
phtfaalate
Chlordane
DDT
Lindane
PCBs
Trichloroethylene
Aerobic
Degradation
Rate (yr1?
lfie
0,48*
•II1
0"
0.041
1.2-
0.063°
0.781
Anaerobic
Degradation
Rate (yr"1)6
Af
lr
0.12"
0*
36*
2.5fc
8.3"
0.00063I1
3.3'
Unsaturated
Zone
Degradation
Rate (yr1)0
1^
.6
0.048
1.1
0
0.004
1.2
0.0063
0.78
Saturated
Zone
Degradation
RateCy^)'
0.8
0.084
0.55
18
1.3
4.8
0.0035
2.0
*"
b Based on microbial degradation rates.
^ biodcsradation rates" Hydrolysis rates used for lindane and
**a*JB
°f *» unsatu«ted »» degradation rates and the anaerobic
e Vaishnav and Babeu (1987).
f Horowitz «/ a/. (1982).
* Coover and Sims (1987).
" Anaerobic rate assumed to equal 25% of aerobic rate
1 Howard et al. (1991).
j Shelton et al. (1984).
fc Castro and Yoshida (1971),
1 Stewart and Chisholm (1971).
m Ellington et al. (1988).
"•Zhang et al. (1982).
0 Fries (1982).
p Anaerobic rate assumed to equal 1 % of aerobic rate
* Dilling et al. (1975).
r Bouwer and McCarthy (1983).
7-45
-------
assumed to undergo significant hydrolysis: since hydrolysis rates are far more accurately
quantifiable than microbial degradation rates, hydrolysis rates are used for these two chemicals.
For the other six organic contaniinarits, 10 percent of the aerobic biodegradation decay rate is
assumed to be appropriate for the unsaturated zone. This decision is based on the observation
that/« tends to decrease with depth in the soil, thereby reducing the amount of suitable substrate
for microbial populations which might degrade these chemicals (O'Connor, 1992).
In the saturated zone, all three degradation processes can occur because some
groundwater is anaerobic and some aerobic. To capture this mix of processes, we have
calculated an arithmetic mean from the aerobic and anaerobic biodegradation decay rates
discussed above. For lindane and itrichloroethylene, the only two chemicals where hydrolysis
is a significant degradation process, estimated anaerobic decay rates are significantly higher than
hydrolysis rates.
For PCBs, it is difficult to assign an anaerobic degradation rate. Highly chlorinated
congeners may be partially degrade very slowly in reducing conditions, but then oxidative
conditions must be established for further degradation to occur. Adequate information on
anaerobic degradation rates cannot be obtained from the scientific literature. We have
conservatively assumed that anaerobic degradation of PCBs occurs at 1 percent of the aerobic
biodegradation rate.
Henry's Law Constant
Henry's Law constants are used to calculate the rate at which organic contaminants
volatilize from sludge. Determining appropriate values for these constants is complicated by the
wide variation in estimates provided by various sources. Table 7-11 shows values taken from
four different sources, along with the value selected for this analysis. Whenever possible, values
are taken from Lyman et al. (1990); otherwise values are taken from: the GEMS data base (U.S.
EPA, 1988a), the PCGEMS data basic (U.S. EPA, 1989d), or the Aquatic Fate Process Data for
Organic Priority Pollutants (U.S. EPA, 1982a). The decision process is as follows: if a value
is published in Lyman et al. (1990), it is used. If not, but if two values from other sources are
similar, the mean of those two values is used. If there is no value in Lyman et al. and no two
values agree, a measured value is chosen in preference to an estimated one. If only estimated,
dissimilar values are available, the value most conservative for groundwater (i.e., the lowest
Henry's Law constant) is chosesn. This last circumstance occurs only for bis(2-
ethylhexyi)phthalate.
The only exception to the decision process described above is for polychlorinated
biphenyls (PCBs), which include a variety of possible congeners with different chemical
characteristics. Anderson and Parker (1990) provide a compilation of non-dimensional Henry's
Law constants for one penta-congener and three hexa-congeners. To derive a representative
Henry's Law Constant for PCBs, th« three values for hexa-congeners were averaged to a single
value which was then averaged with the penta-congener value to obtain the single constant
reported in Table 7-11.
7-46
-------
j>
c
CK
,_, C
| (j
t**
I I
a t
21
!.i
P
PCGEMS<«-b>
(atm-mVmol)
s s
$2222222
o
eo'
X X
O VO
JS 14 x
^J "^^ ^Q
C^ C^ ^
£ £
^ I
2 | 2 22
to
o
o
ci
—
*>>
•a
t
03
_a
o -r
2 '•§
H-S g
|I§
3 « 3
Wl
-------
For all organic contaminants except PCBs, the dimensional estimate of Henry's Law
Constant reported in Table 7-11 has been converted to an equivalent non-dimensional constant
based on an assumed temperature of 15 *C (288K) and the following equation:
H-^-
~ RT
where:
T = temperature (assumed to be 288 K),
R . = Ideal Gas Constant (m3-atm/mol-K),
H = dimensional Henry's law constant (m3-atm/mol), and
H = non-dimensional Henry's Law constant.
Because Anderson and Parker (1990) report non-dimensional values for PCBs at 25 °C, the
average value derived from this source has been adjusted to an equivalent non-dimensional value
at 15 °C,
Diffusion Coefficients
Volatilization of contaminant from a surface impoundment is modeled with a mass
transfer coefficient derived with a two-layer resistance model. Because contaminant must pass
through both the liquid and air to be released into the atmosphere, the overall resistance equals
the sum of the liquid and gas phase resistances,which.aredescribed as the inverse of mass
transport-coefficients for each phase. Methods for calculating these mass transfer coefficients
are selected according to two types: of site characteristics: (1) the ratio of the surface's effective
diameter (or "fetch") to its depth and (2) the local average wind speed. Effective diameter is
defined as the diameter of a circle of area equal to the facility's. The fetch:depth ratio for the
model site is about 80 and is calculated using an effective diameter of 161 m (area = 20,236
m2), and a depth of 2 m (the depth of the liquid layer).
Mass transfer coefficients for the liquid and gas phases are calculated from effective
diameter, the fetckdepth ratio, wind-velocity, the viscosity and density of air, and the estimated
diffusivity of each contaminant in water and air. Default values for the viscosity of air (1. 8x10^
g/cm-sec) and the density of air (1.2x10* g/cm3 at STP) have been taken from Incropera and
DeWitt (1985). Wilke and Lee's method provides estimates for the diffusivity of each
contaminant in air, and Hayduk and JLaudie's method provides estimates for each contaminant's
diffusivity in water (Lyman et al., 1990). The resulting estimates, which are based on a
temperature of 15*C, are listed in Table 7-12.
Molecular Weights
The values presented in Table 7-13 are standard molecular weights for the contaminants
of concern. These weights are used in the vapor loss component of the mass balance program.
7-48
-------
Table 7-12
*
Diffusion Coefficients for Organic Contaminants
Diffusivity in Air Diffusivity in Water
(cmVsec)'
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Chlordane
DDT
Lindane
Polychlorinated biphenyls
Trichloroethylene
9.1 x 10-2
4.6 x 10"2
3.3 x 10-2
4.5 x 10-2
4.1 x 10-2
5.0 x ID"2
-5.7x.10*
8.2 xlO-2
7.8 x 10-*
4.3 x 10^
3.2 x 10-*
3.7 x 10^
3.7 xlO-*
4.5 x 10-6
4.2 x 10*
7.3 x .10^
• Calculated using the Wilke and l.ee method from Chapter 17-4 of Lyman et al. (1990)
based on temperature of 15 °C. --
b Calculated using the Hayduk and Laudie method from Chapter 17-7 of Lyman et al
(1990), based on temperature of 15°C.
7-49
-------
Table 7-13
Molecular Weights for Organic Contaminants
Molecular Weight
Benzene 7g.l
Benzo(a)pyrene 252.3
Bis(2-ethylhexyl)phthalate 390.6
Chlordane 409.8
DDT 354.5
Lindane 290.8
Polychlorinated biphenyls (Aroclor 1254) 325.1
Trichloroethylene 131.4
7-50
-------
7.2.4 Size of Exposed Population
the methods °Pulations surrounding each individual SMA. Based on data
from the National Well Water Association, we determine the number of households served by
L™ of Pth Jt^? ™ ** r n^ntaining «* surface disposal facility in the analytic
firifed bv Sfe SS; ^S vfe; when multiplied by ** avera*e household size «* *«
dmded by the total area of each relevant county, provides an estimate of the average density of
™S± a\UMg ^J^T * «* cou<*y- R* «ie population exposed through the a*
pathway, a total population density was calculated based on the county population and county
^ eXpOS6d m for *» g^^water pathway is taken S
flowre 7"* °r ' CWlBie °n A'W-B-* don of groundwatr
flow Figure /-I depicts the exposed areas for both the groundwater and the air pathways The
potentud for exposure to contaminated groundwater is conservatively assigned such that persons
S?«Jt^ Tam/ ** ^ "ff^ t0 the «««^» «ta-5« the downSn
edge of the site, and persons hving between 1000-3000 m from the site are exposed to the
concentrations estimated at 1000 m down gradient of the site. For the air pathwTTreScfcS
conations of contaminant at the property boundary (150 m from the edge in the^owTwmd
duection) are conservatively applied to the entire population residing in the exposed area. The
j rn i_t~ _ ,^~ _* -—«— . »T uuu laDie /-15 for mononlls, and
and Table 7-17 for surface impoundments.
7.3 BASELINE RISKS
Tables 7-18 through 7-21 provide our estimates for the health risks caused by monofilling
lee Sludge under hasAlino ^nnrii+j,™*. -oi^ Qf hnth cance H h h ft?
and about 140,000 for the groundwater pathway. As can be'seen^SnTthe''^!^
nsks of cancer from facilities represented by the monoffll prototype are expected tote
OUTWIT rrt//^ «««x^*u-KMrtAd.x.1 o ' *• »* £»—»•.<»*• fc*^- w
or cancer exuected every "
riu^f v f ^ ^ -to attnbuted to »»»fc from the groundwater pathway. Cancer
nsks to the highly exposed individual are estimated to be about IxlO4.
nnn, f^ non-carcinogenic **». ™»te 7-19 and 7-20, show that without background sources
enic iofh P RfD' » estmate
arsenic are 17% of the RfD for the average individual and 55% for the HEL These values may
7-51
-------
Figure 7-1
Zones of Human Exjwsure for Groundwater and Air Pathways
7-52
-------
GO
i
o
§
&
03
§
i 1
CO
Ml
a
Q
O. x--
o -C^ c
fi -s -S
si I
a
o
JT
8
O
S M
=
•f
S
B
1
3 «
o
P §
ON, ^
S'
S
-«-
s 2 » 8
g; o g; a
a
1 1
1 I
4) O
« 2
5-
S
I I I
|
.3
o —
53 oe a-
3>3.i
•s ~ s
t
i.™
ID z
-------
<« g
.2 H ~
I
& D 2
.
«N
00
'O
s
~ *•»-<•"«
ill
I
00 00
i » 8
-"
§ 5 =
vo —< —
2! S3 •-< £.. o\ o
S 1-i 3 38 S I
09 O
2 .3
I I I a •« B |
< s.
s
o 2
3 1
S .:
•9 I
o o
a a
-------
.2
i
T3
I
V
'-Jt ' -O
a !
o
*
1
a
<£
3
S :
o
00 «
2
J
2
li
I
l
II
i
l
lI
§
60
a>
5
§•
1
u
-------
I
•o
c
u
I
<£
t^ e
t I
~ o.
I
CO
3
a
1
«s
g
3
O.
fi
BO
.3 .
en
J u|
O" ^^
u
A
tt.
^t- r- n
t*» *o V"i
•* vo v>
qj en -4 ~
co
s
6
o
53
O
s
O
s
(3 55
55 a
§•
co
s*
<£
I i
«
• VD
-------
Table 7-18
Baseline Cancer Risks for Surface Disposal:
Monofffls*
Contaminant
Arsenic
Benzene
Benzo(a)pyrene
Bis(2-ethylhexyl)phthalate
Chlordane
DDT
Lindane
PCBs
Trichloroethylene
Total
AVERAGE INDIVIDUAL RISK"
INDIVIDUAL RISK FOR HET1
Groundwater
0.02
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.02
IxlO-5
3x10-*
Vapor
__
<0.01
<0.01
<0.01
<0.01
<0.01
. <0.01
.. <0.01
<0.01
<0.01
IxlO-10
4xlO-9
Total"
0.02
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.02
1x10*
SxlO4
* These results are based on reasonable worst-case input parameters and assumptions.
b Individual totals may not sum to totals because of independent rounding.
0 Risk for average exposed individual of developing cancer from lifetime of exposure to pollutants from
sludge.
4 Risk for Highly Exposed Individual (HEI).
7-57
-------
8
o
o
W\ 3
U
* I
1!
I
So -S
e
2
£
b
S0 -8
S
5
2
S
III I
a
11
»H ^^
i
~ 60
&<*
Is
.1
2 &
la
S <=>
V V V V
o
V
O (b O
TI —< T!
1 2
.C. X
X
w>
_• ^i 00 M'
o
V
X
oo
O O O O
^•4 T—4 ^H »«H
O\ CS
H
•1 1 's- S.
|ll l|3
•< U U U S S3
00
to.S
u
0>
-------
o
Q
u
,«3
.M
( H
2 <
«
* >
« "fi-l
5 S e
•" o §
Q. C
43 =2
•S3
O O
V3 J3
«S
t*. «
O i-
o
CO
6
§
e
"S
£ •»
§ » e
a .5 a
'"§ 2 6D
A
«
60
S
8
% •§ f J?
w *£; Q "50
0^5000
V V V V V
ooo
*3 73 *3
-------
Table 7-21
Baseline Non-Cancer Health ,Risks*
Monofilling i
Non-Cancer Health Effects
Persons Exceeding RFDC
Persons Crossing Cadmium Threshold51
Persons Crossing PbB Threshold'
Men
Women
Children
Total Crossing PbB Threshold
TOTAL NON-CANCER
Expected Disease Cases from Lead8
Men
Women
Children
TOTAL
Groundwater
0
<0.01
<0.01
<0.01
«3.01 I
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
Vapor
0
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
Total"
0
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
b Individual values may not sum to totals because of independent rounding.
e Number of persons crossing RfD exposure threshold for one or more contaminants because of sludge reuse
or disposal.
* Estimated number of persons with kidney cadmium levels rising above 200 /tg/g. For discussion, refer to
Chapter 2.
e Estimated number of persons with blood lead (PbB) levels crossing health effect thresholds. For discussion
of assumptions and methodology, refer to Chapter 2.
f Number of persons crossing thresholds of lead or cadmium body burden because of exposure to sludge. The
number of persons experiencing adverse non-cancer health effects from theses two contaminants is likely to
be lower.
* Not all of the persons exceeding blood lead levels can be expected to suffer adverse health effects. These
values represent estimates of actual disease cases resulting from sludge re-use or disposal. For discussion of
disease and estimation methodology, refer to Chapter 2. '
7-60
-------
be significant when accounting for background exposure sources. For instance combined
background and monofffl exposure sources yields estimates that exceed the RfD for arsenic by
430% for the average individual iind 470% for the HEI;! however, note that the background
levels alone accounts for most^of this exposure in that the background for arsenic exceeds the
RfD by 400%.
As shown in Table 7-21, special calculations of risks for lead and cadmium suggest that
fewer than one person per hundred years will cross body-burden thresholds or experience
adverse health effects because of exposure through the groundwater pathway from sludge
monofills. Because these estimates are based on several worst-case assumptions, actual risks for
both cancer and non-cancer effects are likely to be lower.
Tables 7-22 through 7-25 provide the corresponding tables for the surface impoundment
PI?°i^J^ ,*t0tal expOSed P°Pulation of a130"! 1,500,000 persons for the vapor pathway
and 170,000 for the groundwater pathway, we expect about one incremental case of cancer
nationally per each twenty years thiis practice continues. As with the monofill prototype more
than 99 percent of the total risk is caused by arsenic through the groundwater pathway. Cancer
risk to the highly exposed individual is estimated to be about 6x10*.
For non-carcinogenic risks, predicted baseline and HEI exposure to all metals, excluding
background intake, is lower than RfD levels. For arsenic, however, without background
exposure, the baseline HEI exposure in a sludge impoundment represents 96 percent of the RfD
(Table 24) and the baseline average exposure is estimated to be about 27 percent (Table ^
This implies that sludges with slightly higher levels of arsenic might lead to exposure levels ii
excess of the risk reference dose. However, we note that these estimates are based on several
worst-case assumptions (e.g. that the well is assumed to be located at the property boundary in
the down-gradient direction; the soiil is assumed to be sandy; and the depth to groundwater is
assumed to be 1 m), so that actual exposure and risks may be likely to be much lower than those
predicted.
As shown in Table 7-25, we do not predict significant health risks from lead and
cadmium, for which fewer than one person per hundred years is expected to cross body burden
thresholds or experience adverse health effects as a result of exposure to lead or cadmium from
surface impoundments.
Table 7-26 presents combined estimates of the monofilling and surface impoundment
prototypes to give total baseline cancer risks from surface disposal. In total, we estimate fewer
than seven incremental cases of cancer might be caused every hundred years as a result of
surface disposal. As mentioned earlier, this estimate is based on several worst case assumptions
so actual risks are likely to be even lower. '
7.3.1 Benefits from Regulatory Controls
Based on an analysis of those 26 facilities in the analytic survey of the NSSS believed
to use surface disposal for their sludge, it appears unlikely that a significant number of facilities
7-61
-------
Table 7-22
Baseline Cancer Cases*:
Surface Impoundments
Contaminant
Arsenic
Benzene
Benzo(a)pyrene
Bis(2)ethylhexylphthalate
Chlordane
DDT
Lindane
PCBs
Trichloroethylene
Total
AVERAGE INDIVIDUAL RISK6
INDIVIDUAL RISK FOR HE?1
Groundwater
0.05
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.05
2xlO-s
6x10-*
Vapor
—
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
5x10-*
IxlO"5
Total"
0.05
<6.01
<0.01
•
-------
s ^
SI
(U 03
e ^
§
,«
jj
3- •»•
S S
RO
§
« e
ea 3
S &
*l
II
0,3
g W
O *->
O ,P
*-. ci
e Exposu
re
M^^
.S §-S
lit
'•ajjr
J M^
P ^ M >
lit
I
1
f
IR
55
V
o
V
T S.
O o
JS X *
°. CS CJ}
co _; in
x x
IS 00
CO
*s 83 ss
S K « 2
000
i—< i—i «—<
CO OO ^^
cs
-------
fi« -a,
|. |
n. "5i)
s
s
t
vo
Ox
c>
V
r*1
'5
o
V
8ft
V-l
o o o o o CD
3 "* • S 3 "* **
ON oo en \e ^ —<
O
v—4
X
o
?™^
X
-H en
en
X
en
X ' X X
in "—' ID
X
en
b^ S^ &^[ ^^
oo *n esi Th
en -^ es; cK
O 0
s-a
S .a
S g
a x
2
-------
Table 7-25
Baseline Non-Cancer Health Risks1
'• Surface Impoundment
•
Non-Cancer Health Effects
Persons Exceeding RFDC
Persons Crossing Cadmium irhresholdd
Persons Crossing PbB Threshold"
Men
Women
Children
Total Crossing PbB Threshold
TOTAL NON-CANCER
Expected Disease Cases from Lead*
Men
Women
Children
TOTAL
Groundwater
0
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
Vapor
0
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
Total"
0
«3.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
* Based on "worst case" input parameters and assumptions.
" Individual values may not sum to totals because of independent rounding.
e Number of persons crossing RfD exposure threshold for one or more contaminants because of sludge reuse
or disposal.
4 Estimated number of persons with kidney cadmium levels rising above 200 ag/g. For discussion refer to
Chapter 2.
e Estimated number of persons with blood lead (PbB) levels crossing health effect thresholds. For discussion
of assumptions and methodology, refer to Chapter 2.
f Number of persons crossing thresholds of lead or cadmium body burden because of exposure to sludge. The
number of persons experiencing adverse non-cancer health effects from theses two contaminants is likely to
be lower. ;
* Not all of the persons exceeding blood lead levels can be expected to suffer adverse health effects. These
values represent estimates of actual disease cases resulting from sludge re-use or disposal. For discussion of
disease and estimation methodology, refer to Chapter 2.
7-65
-------
Table 7-26
«
Baseline Cancer Cases*:
Total Surface Disposal
Contaminant
Arsenic
Benzene
Benzo(a)pyrene
Bis(2)ethylhexylphthalate
chlordane
DDT
Lindane
PCBs
Trichloroethylene
Total
AVERAGE INDIVIDUAL RISK6
INDIVIDUAL RISK FOR HEId
Groundwater
0.07
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.07
2xlOJ
6x10*
Vapor
1 —
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0-01
SxlO"8
IxlO-5
Total"
0.07
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.07
2x10-*
6x10*
1 These results are based on reasonable worst-case input parameters and assumptions.
* Individual totals may not sum to totals because of independent rounding.
° Risk for average exposed individual of developing cancer from lifetime exposure to pollutants from sludge.
* Risk for Highly Exposed Individual (HH).
7-66
-------
using tins method will be forced to change to other practices as a result of the regulation
Although the additional management practices required by the regulation are likely to reduce
expected human exposure and risk, their effects are not directly quantifiable by methods used
for this analysis. Nevertheless, the potential reduction in risk likely to be achieved by the
regulation should not exceed our estimates of risk for baseline conditions. Therefore we
estimate that the potential benefit of regulation falls within the range of zero to our estimate of
baseline nsk.
7.3.2 Uncertainties and Limitations
This analysis is based on numerous simplifying assumptions (almost all of them
conservative). To the extent that actual conditions at individual facilities differ from those
assumed, true exposure and risk will differ from the estimates provided here. For example if
the local depth to groundwater exajeds 1 m, or if the hydraulic conductivity of the local mil
medium is less than that of sand, actual contamination of groundwater beneath a surface disposal
facility is likely to be lower than that calculated for this analysis. Similarly, concentrations in
groundwater at distances greater than 150 m in directions other than downgradient are likely to
be lower than those calculated for this analysis. On the other hand, a non-homogeneous or
fractured medium beneath a surface disposal facility might lead to the contamination of
groundwater at higher levels than those predicted by VADOFT and AT123D. Moreover, if the
number of persons drinking groundwater near a surface disposal facility exceeds the average
density calculated for the county as a whole, risks at these sites may be underestimated. Finally
if the density of human populations or concentrations of contaminants in sludge for other surface
disposal facilities differs systematically from those predicted based on the analytic component
of the NSSS, risks might be underestimated. Nevertheless, estimates of exposure and risk
derived in this analysis are believed to be conservative, and unlikely to underestimate true risks
7-67
-------
8. REFERENCES
Abt Associates Inc. 1989. Characterization of Surface Disposal for Wastewater Sludge.
Memorandum to Alan'Rubin, U.S. EPA Office of Water Regulations and Standards.
March 27.
Anderson, M.A. and J.C. Parker. 1990. Sensitivity of organic contaminant transport and
persistence models to Henry's Law constants: case of polychlorinated biphenyls.
Water. Air, and Soil Pollution. 50: 1-18.
Angle, C.R., and M.S. Mclntire. 1979. Environmental lead and children: the Omaha
study. Journal of Toxicol. JEnviron. Health.. 5:855-870.
Baes, C.F. m, Shor, R.W., Sharp, R.D., and .AX. Sjoreen. 1985. "Two Agricultural
Production Data Libraries for Risk Assessment Models," Journal of Environmental
Quality. 14(4).
Baxter, J.C., D. Johnson, E. Klienholz, W.D. Burge, and W.N. Cramer. 1983. Effects on
cattle from exposure to sewage sludge. PB83-170589. EPA-600/2-83-012.
Binder, S., D. Sokal and D. Maughan. 1986. "Estimating Soil Ingestion: The Use of
Tracer Elements in Estimating the Amount of Soil ingested by Young Children."
Arch. Environ. Health. 41:341-345.
Booz-AUen and Hamilton. 1982. Model Plant Risk Analysis Exercise: Report on Criteria
for Defining Model Plant Selection. October.
Bouwer, E.J. and P.L. McCarthy. 1983, Transformations of 1- and 2- carbon halogenated
aliphatic organic compounds under methanogenic conditions. Applied and
Environmental Microbiology. 45(4): 1286-1294.
Bowers, J.F. et al. 1980. Industrial Source Complex (ISC) Dispersion Model User's Guide
(Vol. 1). PB80-133044. U.S. EPA. Research Triangle Park. NC.
Branson, D.R., et al. 1985. Biocwncentration kinetics of 2,3,7,8-tetracMorodibenzo-p-
Dioxin in rainbow trout. Einvironmental Toxicology and Chemistry. 4:779-788.
Brauer, H., and Y.B.G. Varmai 1981. Air Pollution Control Equipment Springer-Verlag,
Berlin Heidelberg. New York.
Carsel, R.F. and R.S. Parrish. 1988. Developing joint probability distributions of soil
water retention characteristics. Water Resources Research. 24(5): 755-769.
Castro, T.F. and T. Yoshida. 1971. Degradation of organochlorine insecticides in flooded
soils in the Philippines. J. Agr. FoodChem.. 19(6): 1168-1170.
8-1
-------
Effect of airborne lead on blood lead. Atmospheric Environment
Chancy, R.L., R.J.F. Bruins;tD.E. Baker, R.F. Korcak, J.E. Smith, Jr., and D.W. Cole.
1987. Transfer of sludge-aipplied trace elements to the food-chain, pp. 67-99. In
AX. Page, T.J. Logan, and J.A. Ryan (eds). Land Application of Sludge -Pnnrf
Chain Implications. Lewis Publishers Inc., Chelsea, MI.
Chaney, R. 1992. U.S. Department of Agriculture. Personal Communication.
Cohen, J. 1987a. Respiratory Deposition and Absorption of Lead Particles. M^mn^nHnn.
to Fred Miller and Ted Martorien, Inhalation Toxicology Divisions, U.S. EPA, Office
of Air Quality Planning and Standards, Ambient Standards Branch, Research Triansle
Paric, NC, October 7, 1987. *
Cohen, J. 1987b. Risk Assessment for Municipal S1ndffft nigpn^i Memorandum to Betsy
Southerland, U:S. EPA Office of Water Regulations and Standards, December 29.
Cohen, J. 1988. U.S. EPA Office of Air Quality Planning and Standards, Ambient
Standards Branch, Research Triangle Park, NC. Personal communication. August.
Coover, M.P. and R.C. Sims. 1987. The effect of temperature on polycyclic aromatic
hydrocarbon persistence in sm unacclimated agricultural soil. Hazardous Waste and
Grumpier, E. 1991. U.S. EPA, Office of Air Quality, Planning and Standards. Personal
Communication.
Davis, M., and Svendsgaard, D.J. 1987. Lead and Child Development, Nature Vol 329
September, 1987.
Dewling, R.T., Manganeffi, R.M, and G.T. Baer Jr. 1980. Fate and behavior of selected
heavy metals in incinerated sludge. Journal WPCF Vol. 52, No 10. Northwest
Bergen Sewer Authority, Wiildwick, NJ.
DietricL, K.N., Krafft; K.M., Shukla, R., Bornschein, R.L., Succop, P.A. 1986. The
neurobehavioral effects of prenatal and early postnatal lead exposure. In: Schroeder,
S-R-> «*• Toxic Substances and Marital Retardation: Neurobehavioral Toxicology and
Teratolpgy. Washington DC: AAMD monograph series.
Dilling, W.L., N.B. TefertiUer, and G.J. Kallos. 1975. Evaporation rates and reactivities
of methylene chloride, chloroform, 1,1,1-trichloroethane, trichloroethylene, -
tetrachloroethylene, and other chlorinated compounds in dilute aqueous solutions.
' Environmental Science & Technology 9(9): 833-838.
8-2
-------
Donnelly, J.R., and Quach. 1985. Joy/Niro Spray Dryer Absorption System. Paper
presented at: Acid Gas and Dioxin Control for Waste to Energy Facilities Conference,
Washington, DC. November 25-26.
*
Ellington, J.J., F.E. Stancil, W.D,, Payne, and C.D. Trusty. 1988. Measurement of
Hydrolysis Rate Constants for Evaluation of Hazardous Waste Land Disposal:
Volume 3. Data on 70 Chemicals. EPA-600/3-88-028. NTIS PB88-234 042/AS, as
cited in Handbook of P.nvirnnrnental Degradation Rates.
Environment Canada. 1986. The National Incinerator Testing and Evaluation Program: Air
Pollution Control Technology. Summary Report, EPA 3/UP/2.
Environmental Science and Engineering. 1985. Exposure to Airborne Contaminants
Released from Land Disposal Facilities - A Proposed Methodology. Prepared by
Environmental Science and Engineering, Gainsville, FL, for the U.S. EPA Office of
Solid Wastes. Washington DC.
Fairbanks, B.C. and G.A. O'Connor. 1984. Effect of sewage sludge on the adsorption of
polychlorinated biphenyls by three New Mexico soils. Journal of Environmental
Quality. 13:297-300.
Freeze, R.A. and J.A. Cherry. 1979. Groundwater. Englewood Cliffs, NJ: Prentice-Hall.
Fries, George F. 1982, Potential polychlorinated biphenyl residues in animal products from
application of contaminated sewage sludge to land. Journal of Environmental Quality.
11(1): 14-20.
Fulton, M., Thomson, G., Hunter, R., Raab, G., Laxen, D., and Hepburn, W. 1987,
Influence of blood lead on the ability and attainment of children in Edinburgh, The
Lancet. May 30, pp. 12214226.
Gerritse, R.G., R. Vriesema, J.W. Dalenberg, and H.P. De Roos. 1982. Effect of sewage
sludge on trace element mobility in soils. Journal of Environmental Polity 11(3)'
359-364.
Gerstle, R.W., and Albrinck, D.N. 1982. Atmospheric emissions of metals from sewage
sludge incineration. Journal of the Air Pollution Control Association. Vol. 32, No
11, Nov. 1982.
Goroll, A.H. and L.A. May, eds. 1981 Primary Medicine Care. Management of
Hypertension, prepared by.E.E. Slater. J.B. Lippencott Co., Philadelphia, Pa.
Griffin, R.A. 1991. Director, Environmental Institute for Waste Management Studies,
Dept. of Chemical Engineering, University of Alabama. Personal Communication.
8-3
-------
iZ^h Lan RX- Schmouder' N-G- Goldstein, and L.C. Harian 1985
Blood Lead and Blood Pressure: Relationship in the Adolescent and Adult U S
Population, Journal of the American Medial A^iari™, 253:530-534.
Hartunian N.S., C.N Smart, and M.S. Thompson 1981. The Incidence an
of Major Health Impairments. Lexington Books, Lexington, MA.
Hasset J J, W.L. Banwart, and R.A. Griffin. 1983. Environment aiiH Snin ,..
Characterization, Treatment,, and Disposal, Edited by Francis, C.W. S I Auerbach
and V.A. Jacobs. Butterwortfa Publishers, Woburn MA. pp. 161-175. AUerbaCh'
' 1987' MResidual m«** of ewa«« uge on
1 Accumulatlon of Heavy Metals." Journal of *
Hem, J D 1985 Significance of properties and constituents reported in water analysis -fa-
Study and Merpretatmn of the Chemical rhan^ric^ »f Natural Water. 3^
' WatoMIW1y PaPer 2254- Reston. VA: USGS Department of Interior.
R> Shelton' C:P- C?meU' ^ J.M. Ifedje. 1982. Anaerobic degradation
d^Sl
Howard, P.H. 1991. Handbook of Environmental Fa<:e and Exposure Data for
Chemicals. Vol m, PesticMg. Lewis Publishers, Inc. Chelsea, MI.
w';' W'R Jarvis' W'M- M&y^ ™dE- M. Michalenko. 1991
Handbook of Environmental IDeeradation Ratev T^i, Publishers, Inc. Chelsea, ML
Hwang^S.T.^982. Toxic emission from land disposal facilities. Environmental Pmpm«
Hwang^S.T.^1985. Model prediction of voktile emissions, Bivironment.1 Pmp^ 4(2):
Hwai:g, S.T. and . Falco.' 1986. Estimation of multimedia exposures related to hazardous
S; Newark0: "' ^ ^^ " ^^^ ^ironment. Plenum
Incropera P.P. and D.P DeWitt. 1985. Fundamentals of Heat and M.« Trangfef John
Wiley and Sons, Inc. New York. -
- • •- ^ . .~— .„. . ,_„,.'- , ,, >
Javitz, H. 1980. .Sea-Food Consumption Data Analysis. SRI International. Menlo Park,
8-4
-------
Levy, R.I., et al. 1984. The decline in coronary heart disease mortality: status and
perspective on the role of cholesterol. American Journal of Cardiology. August 27:
544.
Lyman, W.J., W.F. Reehl and D.H. Rosenblatt. 1990. Handbook of Chemical Property
Estimation Methods: Envimnmanfal Behavior of Organic Compounds. American
Chemical Society. Washington D.C.
MacKay, D. and PJ. Leinonen. 1975. Rate of evaporation of low-solubility contaminants
from water bodies to atmosphere. Environmental Science & Technology 9(13)-
1178-1180. !
MacKay, D. and A. Yeun. 1983. Mass transfer coefficient correlations for volatilization of
organic solutes from water. :Bnvironmental Science & Technology. 17(4): 423-429.
Mahaffey K.R., Rosen, J.F., Chesney, R.W., et al. 1982. Association between age, blood
lead concentration, and serum 1,25 dihydroxycholecalciferol levels in children Am
J. Clin. Nutr. 35: 1327-1331,, 1982.
Marcus, A.H. 1988. Applied Statistics and Computer Applications Section, Battelle Inc.
Memorandum to Jeff Cohen, U.S. EPA Office of Air Quality Planning and Standards,
Ambient Standards Branch, Research Triangle Park, NC. February 11, 1988.
Marcus, _A.H. 1989. Relationship Bstween Childhood Blood Lead and Lead in Water or
Diet. Battelle7Columbus Division,"Arlington, M.D.
McGee and Gordon 1976. The Results of the Framingham Study Applied to Four Other
U.S.-based Epidemiologic Stutdies of Coronary Heart Disease. The Framingham
Study: An Epidemiological Investigation of Cardiovascular Disease. Section 31,
April.
National Academy of Sciences (NAS). 1983. Risk Assessment and Management:
Framework for Decision Making. Washington, D.C.
National Center for Health Statistics. National Hospital Discharge Survey, various issues.
National Center for Health Statistics 1981. Plan and Operation of the Second National Health
and Nutrition Examination Survey 1976-1980. Vital and Health Statistics, Series 1
No. 15. '
National Gardening Survey Association, Inc. 1987. National Gardening Survey 1986-
1987.
National Institute of Occupational Safety and Health (NIOSH). 1979. Registry of Toxics
Effects of Chemical Substances.
8-5
-------
Nationd Well Water Association. 1985. Well Fax: Statistical Data for the States and
Counties and Zip Codes of the United States.
O'Connor, G. 1992. Professor and Chairman, Soil and Water Science Dept, University of
Florida. Personal Communication.
j£.
Pennin^onj. A^T 1983. Revisk-n of the Total Diet Study food list and diets, JLAm^Die
ASSOC. oZ.lOO-J. 13,
' H,'^' S^waf ' J*r ^ldis' **d W-R- Harlan 1985. The relationship between
blood lead levels and blood pressure and its cardiovascular risk implications
American Journal of Epifteminlngy 121:246-258.
Pocock, S J A.GL Shaper, D. Ashby, T. Deleves. 1985. Blood lead and blood pressure in
middle aged men. In: Lekkas, T.D. (ed.), International Conference on Heavy Metals
in the Environment. Athens, Greece. September.
Pooling Project Research Group 1978. Relationship of blood pressure, serum cholesterol
smoking habit, relative weight and ECG abnormalities to incidence of major coronary
events: final report of the pooling project. Journal of Chmnin in—. « 3{. ™L
Rabinowitz, M Bellinger, D., Leviton, A., Needleman, H., and Schoenbaum, S. 108?"
Chapter 2: Pregnancy hypertension, blood pressure during labor, and blood lead
levels. Hyr>ertension,_VpllO.,No.4, October.
Rabinowitz, M.B., Wetherffl, G.W., Kopple, J.D. 1977. Magnitude of lead intake from
respiration by normal man. J. Lab. Clin. MeH 90:238-248.
Roels, H. A . J.P. Buchet RR. lawwrerys, P. Bruaux, F. Claeys-Thoreau, A. LaFontaine
and G. Verduxn. 1980. Exposure to lead by the oral and pulmonary routes of
chd^en living in the vicinity of a primary lead smelter. Environmental E^Tr^ 22:
Rozman, K., T. Rozman, H. Grekn, IJ. Nieman, and G.S. Smith. 1982 Use of
aliphatic hydrocarbons in feed to decrease body burdens of lipophilic toxicants in
livestock. J. ACT. Food rimm ^Qg-inn ^^F
Ryan,J.A. 1991. "ORD review ol' -Paradigm for soU risk assessment: A soa scientist
perspective' Data tables. " Memorandum to Burnell Vincent, U.S. EPA Office of
Research and Development, November 8.
Scheffler and Rossiter, eds. 1984, Advances in H^th Economic atld Health ervs
Researcb, Vol 5. "Hospital Cost Inflation, Health Insurance and Market Incentives "
prepared by A.K. Taylor. JAI Press, Inc. Greenwich, CT
8-6
-------
Schwartz, J., Angle, C., and Pitcher, H. 1986a. Relationship between childhood blood lead «*" \
levels and stature. Pediatrics. 77:281-288. \_. J
Schwartz, J. 1986b. NHANES H Blood Pressure .Analysis [memorandum to Lester Grant].
Washington DC: U.S.* EPA, Office of Policy Analysis. Available for inspection at:
U.S. EPA Central Docket Section, Washington, DC; docket nos. ECAO-CD-81-2
UA.C.9 and ECAO-CD-81-2 UA.C.ll.
Schwartz, J., Otto, D. 1987. Blood lead, hearing thresholds, and neurobehavioral
development in children and youths. Arch. Env. Health, 423 153-160.
Schwartz, J. 1988. The relationship between blood lead and blood pressure in the
NHANES H Survey. Env. Health Persp. Vol. 78, pp 15-22.
Schwartz, J., Landrigan, P.J., Feldman, R.G., Silbergeld, E.K., Baker, E.L., and von
Lindern, I.H. 1988. Threshold effect of lead-induced peripheral neuropathy J
Pediatr.. 112:12-17. ~
Schwartz, J. 1990. I-ead, blood pressure, and cardiovascular disease in men and women.
Environ. Health Perspect.. in press.
Shelton, D.R., S.A. Boyd, and J,,M. tiedje, 1984. Anaerobic biodegradation of phthalic
acid esters in sludge. Environmental Science & Technology. 18(2): 93-97.
Shen, T.T. 1982. Estimation of organic compound emissions from waste lagoons. Journal
of the Air Pollution Control Association. 32(1): 79-82.
Shurtleff, D. 1974. Some Characteristics Related to the Incidence of Cardiovascular
Disease and Death. The Framingham Study: An Epidemiological Investigation of
Cardiovascular Disease. Siection 30, February. !
Silbergeld, E.K., J. Schwartz and! K. Mahaffey. 1988. Lead and osteoporosis: mobilization
of lead from bone in postmenopausal women. Environmental Research. 47, 79-94.
Sledge, D.J. 1987. Size distribution of lead particles at major lead stationary sources:
.source sampling and ambient air monitoring. Memo to John Haines, Chief, ASB
Section B. Office of Air <3uality Planning and Standards. U.S. Environmental
Protection Agency. Research Triangle Park, NC. 23 September, 1987.
Springer, C., P.D. Lunney and K.T. Valsaraj. 1984. Emission of Hazardous Chemicals
from Surface and Near Surface Impoundments to Air. U.S. EPA, Solid and
Hazardous Waste Research ^Division, Cincinnati, OH. Project Number 808161-02.
8-7
-------
' ^ D' ^Wsh0lln- 197L ^S-^rm persistence of BHC, DDT and
chlordane in a sandy loam clay. Cana 1988
U.S. Bureau of the Census. 1991. State and M*m AIB> rtaf. P^V- ^ ^itint1
"Washington, DC, 1991. -
U.S. Bureau of the Census. 1992. Statistical Ah.^,* of tfae unfa* s«^. tooo
^'^; Washington, DC, 1992. '
——. ~,, . ,1*. .. ^ ~ —'» -—•' -^~
&-• .=--->S5l.'.'; ."'• " .... r'-.-.f.--*^
U.S.Dep^rtmentofAgricuitore(IJSDA). 1982. Food and Nutrition Services
Ma"s, Prepared by Schnittker Associates, Washington, D.C.,
U.S. Department of Agriculture (USDA). 1987. Summary Report: National Resources
Inventory. Statistical Bulletin Number 790. Soil Conservation Servke
U.S. Dep«tinent of Agriculture (USDA). 1989. Agricultural Stati^, tQ«o Washington,
U.S. Department of Health and Human Services 1984, Analysis of Birth Cohort of 1983
National Center for Health Statistics. '
U'S"
«^ SerViCeS) ^mc Hedth Service' National Center for
rf- w pl^iure Ievels fa p6180118 18-74 y^ of «* m l976-
nVi^S S ^ MoodpWMuw from 1960-1980 in the United States, Vkal and
Health Stafasti^ Series 11, No. 234. -
U 1977' ^iroimiental Assessment of Subsurface Disposal of Municipal
Wastewater Sludge. Interim Report. Report No. EPA-530/SW-547. Cincinnati OH-
Municipal Environmental Research Laboratory (MERL).
8-8
-------
U.S. EPA. 1978. Process Design Manual for Municipal Sludge Landfills. Office of Solid
Waste. EPA-625/1-78-010/SW-705. ( )
U.S. EPA. 1982a. Aquatic Fate Process Data for Organic Priority Pollutants. Office of
Water Regulations and Standards, Washington D.C. EPA-440/4-81-014.
U.S. EPA. 1982b. Determination of Water-Sediment Partition Coefficients for Priority
Heavy Metals. Submitted by HydroQual, Inc. for the U.S. EPA Environmental
Research Laboratory, Grosse He, MI.
U.S. EPA. 1983a. Rapid Assessment of Potential Ground Water Contamination Under
Emergency Response Conditions. EPA 600/8-83-030.
U.S. EPA. 1983b. Process Design Manual: Land Application of Municipal Sludge. Office
of Research and Development, Municipal Environmental Research Laboratory,
Cincinnati, OH. EPA-625/1-83-016.
U.S. EPA. 1984a. Risk Assessment and Management: Framework for Decision Making.
EPA 600/9-85-002,
U.S. EPA. 1985a. Cadmium Contamination of the Environment: An Assessment of
Nationwide Risk, Office of Water Regulations and Standards. EPA-440/4-85-023,
February.
U.S. EPA. 1985b. Technical Support Document for Development of Guidelines on V-
Hydrologic Criterion for Hazardous Waste Management Facility Location. Draft.
U.S. EPA. 198Sc. Costs and Benefits of Reducing Lead in Gasoline: Final Regulatory
Impact Analysis. Prepared by Economic Analysis Division, Office of Policy
Analysis, February.
U.S. EPA. 1985d. DRASTIC: A Standardized System for Evaluating Groundwater
Pollution Potential Using Hydrogeologic Settings. EPA-600/2-85/018. Ada, OK.
U.S. EPA. 1985e. Health Assessment Document for Polychlorinated Dibenzo-p-Dioxins.
Environmental Criteria and Assessment Office, Cincinnati, OH. EPA 600/8-84-014F.
NTIS PB 86-122546.
U.S. EPA. 1985f. Summary of Environmental Profiles and Hazard Indices for Constituents
of Municipal Sludge: Methods and Results. Office of Water Regulations and
Standards. Wastewater Solids Criteria Branch, July.
U.S. EPA. 1986a. Guidelines for Carcinogen Assessment; Guidelines for Estimating
Exposure; Guidelines for Mutagenicity Risk Assessment; Guidelines for Health
Assessment of Suspect Developmental Toxicants; Guidelines for Health Risk
Assessment of Chemical Mixtures. Federal Register. Vol. 51, No. 185.
8-9
-------
U.S. EPA. 1986b. Guidelines cm Air Quality Models (Revised)
EPA/OAQPS-450/2-78-027R. V«vij«yf
U.S. EPA. 1986c.^Ak Quality Criteria for Lead, Vols. I-IV, Environmental Criteria and
Assessment Office, Cincinnati, OH, EPA 600/8-83/028dF. June. ™
U'S' ^ 198^' ^ftog Lead in Drinking Water: A Benefit Analysis. Prepared by
Office of Policy Planning and Evaluation, Draft Final Report, December
U.S. EPA.. 1986e Research and Development of Ri.sk Assessment Methodology for
Mumproal Sludge Incineration, Environmental Criteria and Assessment Office
" WHIMS,
Cincinnati, OH, March.
U'S' ^ J986J; Estimated National Occurrence and Exposure to Copper in Public
U.S. EPA • 1986h. Research and. Development: Development of Risk Assessment
Methodology for Municipal Sludge ^drilling. Prepared by Environmental Criteria
for
U.S. EPA. 1986L Census of State and Territorial Subtitle D Non-Hazardous Waste
Programs. EPA Office of Solid Waste. EPA-530-SW-86-039.
U.S. EPA. 1986j. Industrial Source Complex (ISC) Dispersion Model User's Guide.-*
Second Edition, U.S. EPA, EPA 450/4-86-005a and 005b. Research Triangle Park,
•iN • \^«
U.S. EPA 1986k. Cross-media Tmnacts of TTHlfaAten and Dispel n ..»o
Sludggs. Regulatory Integration Division, Office of Policy Analysis, October.
Coos-,
U.S. EPA. 1987c. Methodology for Valuing Health Risks of Ambient Lead Exposure
Report prepared by MathTech Inc. for. the Ambient Standards Branch, Office of Air
Quality Planning and Standsirds, U.S. EPA. Contract No. 68-02-4323
8-10
-------
U.S. EPA. 1987d. Options Section Memorandum for. Sewage Sludge Technical
Regulations (40 CFR Part 303).
U.S. EPA. 1987e. Report to Congress: Municipal Wastewater Lagoon Study, Volumes I
and H. Office of Municii>al PoUution Control. (No. 01A0005332)
U.S. EPA. 1987f. Hazardous Waste Treatment, Storage, and Disposal Facilities (TSDF) -
Air Emissions Models. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. EPA-450/3-87-026.
U.S. EPA. • 1987g. Descriptive Statistics on Contaminants in Municipal Sludge Based on the
EPA 40-POTW Study. Analysis and Evaluation Division. Draft Final Report. May.
U.S. EPA. 1988a. Graphical Exposure Modeling System (GEMS) EPA Mainframe. User's
Guide. Prepared by General Sciences Corporation for the Offices of Pesticides and
Toxic Substances. Contract No. 68-02-4281.
U.S. EPA. 1988b. Emissions of Metals and Organics from Four Municipal Wastewater
Sludge Incinerators: Preliminary Data. Prepared by Radian Corp., Research
Triangle Park, NC. May.
U.S. EPA. 1988c. Regulatory Impact Analysis of the Proposed Regulations for Sewage
Sludge Use and Disposal, Office of Water, Economic Analysis Division, Draft
Report, September.
-...."-. . . . . . - . ' • :
U.S. EPA. 1988d. Technical Support Document for the Risk Assessment Methodology for
Land Application and Distribution and Marketing of Municipal Sludge, Office of
Water Regulations Standards, Draft Report, September.
U.S. EPA. 1989a. National Sewage-Sludge Survey. U.S. EPA Office of Water. Office of
Science and Technology.
U.S. EPA. 1989b. Risk of Unsaturated/Saturated Transport and Transformation Interactions
for Chemical Concentrations (RUSTIC), Volume!: Theory and Code Verification.
Prepared by Woodward Clyde Consultants, HydroGeologic, and AQUA TERRA
Consultants for theOffice of Research and Development, Environmental Research
Laboratory, Athens, GA. Contract No. 68-03-6304.
U.S. EPA. 1989c. Risk of Unsaturated/Saturated Transport and Transformation of
Chemical Concentrations (RUSTIC), Volume n: User's Guide. Environmental
Research Laboratory, Athens GA. EPA/600/3-89/048b.
U.S. EPA. 1989d. PC-GEMS Database, .JUser's Guide, Release 1.0. Prepared by General
Sciences Corporation for the Office of Pesticides and Toxic Substances. Contract
NO. 68024281.
i
^
8-11
-------
( U.S. EPA 1989e. Assessing Human Health Risks from Chemically Contaminated Fish and
Shellfish: A Guidance Manual; EPA-503/8-89-002.
U.S. EPA. 1989f. Background 13ocument for the Surface Impoundment Modeling System
(SIMS). Control Technology Center. Research Triangle Park, NC. EPA/600-6-89-
U.S. EPA. 1989g. Human Health Risk Assessment For Municipal Sludge Disposal:
Benefits of Alternative Regulatory Options. Prepared for the Office of Water
Regulations and Standards by Abt Associates, February 1989.
U.S. EPA. 1989h. Development of Risk Assessment Methodology for Land Application
and Distribution and Mariceting of Munic^al Sludge. Office of Research and
Development. Cincinnati; EPA/600/6-89/001.
U.S. EPA. 1989i. Interim procedures for estimating risks associated with exposures to
mixtures of chlorinated dibenzo-p-dioxins and -dibenzofurans (CDDs and CDFs) and
1989 update. U.S. EPA, Risk Assessment Forum^ Washington, DC. EPA/625/3-
89/016.
U.S. EPA. 1990a. Summary of U.S. Geological Survey Drainage Basin Areas and Flows
Prepared for the U.S. EPA Office of Solid Waste by Tetratech Inc
r~~ '" •'" ' '- ' •'- -••- ; • ' ••• .- ••
I j U.S. EPA. 1990b. Guidance on: Assessment:and Control of Bioconcentratable
Contaminants in Surface Waters. DRAFT.
U.S. EPA. 1990c. Development of Risk Assessment Methodology for Surface Disposal of
Municipal Sludge. Prepared by Abt Associates Inc. for the Environmental Criteria
Assessment Office, Office of Research and Development. Cincinnati. ECAChCIN-
U.S. EPA. 1990d. Final Regulatory Impact Analysis of National Primary Drinking Water
Regulations for Lead and Copper. Review Draft prepared by Wade Miller
Associates, Inc. and Abt Associates, Inc. for the Office of Drinking Water, U S
EPA. November 9, 1990. Contract No. 68-CO-0069.
U.S. EPA. 1991. Health Effects Assessment Summary Tables: Second Quarter
Supplement, FY 1991. Prepared by Paul Goetchius, Chemical Hazard Assessment
Division, Syracuse Research Corporation for the Environmental Criteria and
Assessment Office, Cincinnati, OH. SRC TR-91-022.
U.S. EPA. 1992a. STORET Datibase. Data obtained from Louis Holeman, U.S. EPA
Office of Water, Assessment and Watershed Protection Division. : 7
U.S. EPA. 1992b. Guidance on ]BLisk Characterization for Risk Managers and Risk
f / Assessors. Memorandum from F. Henry Habicht H, Deputy Administrator.
8-12
-------
U.S. EPA. 1992c. Technical Sujjport Document for Land Application of Sludge. Office of
Water, Office of Science acid Technology.
U.S. Geologic Survey. 1985. National Water Survey - 1985. Washington, DC.
Vaishnav, D.D. and L. Babeu. 1987. Comparison of occurrence and rates of chemical
biodegradation in natural waters. Bull. .Environ. Contam. Toxicol.. 39: 237-44, as
cited in Handbook of Environmental Degradation Rates.
Vancil M. A., and D. M. White. 1988. Assessment of Add-On Control Technology
Performance for New Municipal Waste Combustors. Prepared for U.S. EPA by
Radian Corporation, EPA (Contract No. 68-02-4378. June 7.
Vanoni, Vita A. (editor). 1975. Sedimentation Engineering. Prepared by the ASCE Task
Committee for the preparation of the manual on sedimentation of the Sedimentation
Committee of the Hydraulics Division. NY. NY.
Wallsten and Whitfield 1986. Assessing the Risks to Young Children of Three Effects
Associated with Elevated Blood Lead Levels. Argonne National Laboratory,
December.
Yankel, A.J., I.H. von Lindern, aind S.D. Walter. 1977. The Silver Valley lead study: the
relationship between childhood blood lead levels and environmental exposure. Journal /""""
of the Air Pollution Control Association. 27:763-767. (
Yeh, G.T. 1981. AT123D: Analytical Transport One-, Two-, and Three Dimensional
Simulation of Waste Transport in the Aquifer System. Oak Ridge National
Laboratory, Environmental. Sciences Division. Publication No. 1439. March.
Zhang, S., Q. An, Z. Gu, and X. Ma. 1982. Degradation of BHC in soil. Huaniing Kexue.
3:1-3, as cited in Handbook of Environmental Degradation Rates.
8-13
-------
APPENDIX A
Partitioning of Contaminant Among Air, Water, and Solids in Soil
Calculations used to derive criteria for groundwater, surface water, and air pathways are
based on the assumption that equffibrium is maintained between concentrations of contaminant
in the air-filled pore space, tins .water-filled pore-space, and the. solid particles of soil.
Equilibrium partitioning between 'dissolved and gaseous phases is described by Henry's Law
constants; partitioning between adsorbed and dissolved phases is described by soil-water partition
coefficients. From these assumptions and the definitions of concentration are derived the
equations used to describe partitioning.
Define:
C, =
Q, =
C. =
Q
Mathematically:
and:
where:
M,,
M,
v.
Vt
v.
concentration of adsorbed contaminant on solid soil particles (kg/kg),
concentration of dissolved contaminant in water-filled pore space of soil
(kg/m3),
concentration of contaminant in air-filled pore space of soil (kg/m3), and
total concentration of contaminant in soil (kg/m3).
C_ ew
„ =
" V
Tvr
. **« * Me» * Mca
r. * vw + vs
mass of adsorbed contaminant (kg),
mass of soil (kg),
mass Of dissolved contaminant (kg),
volume of water in soil (m3),
mass of gaseous contaminant (leg),
volume of air in soil (m3),
total mass of contaminant in soil (kg),
total volume of soil (m3), and
volume of solids in soil (m3).
A-l
-------
The equilibrium distribution coefficient (KD, in mVkg) between adsorbed and dissolved phases
can be defined as:
The dimensionless Henry's Law constant (6) describing the partitioning between gaseous and
dissolved phases is defined as:
The bulk density of soil (JBD, in kg/m3) is defined as:
BD = M, / Vt
The air-filled porosity of soil (0J is defined as:
«.-*"./*;
water-filled porosity (0W) is defined as:
and the total porosity of soil (0,) is defined as:
V. = e.*e O
The above definitions can be combined to yield:
<"„ H H
and:
Ct
-1 = BD KD + 6W +
~ C'w "
and:
KD HJ
These relationships are used throughout the calculations used to derive criteria. Where dry-
weight concentrations of contaminiint in sludge or soil are involved, the equations are modified
slightly, based on the definition:
where:
A-2
-------
"dw
_ .. _.
*" " Ms Vt BD BD
dry-weight'concentration of contaminant in soil.
A-3
-------
APPENDIX B
Derivation of Finst-Order Coefficient for Losses to Leaching
U.S. EPA (1989f) provides an equation for computing a first-order loss rate to leaching
of contaminant from treated soil:
• BD KD
where:
= first-order loss rate coefficient for leaching (yrl),
NR = annual rechairge rate (m/yr),
BD = bulk density of soil (kg/m3),
KD = soil-water distribution coefficient for contaminant (m3/kg),and
dj = depth of incorporation for sludge (m).
This appendix describes a modified version of that equation.
The basic strategy for deriving a coefficient for first-order loss to leaching is to estimate
the mass of contaminant expected to be lost each year and divide by the available mass of
contaminant. The mass of contaminant that will be lost to leaching in any interval of time can
be described by the volume of water percolating through the treated soil multiplied by the
average concentration of contaminant in that water:
FA^ = NR C^ 1.0,000
where:
FAgw = flux of leached contaminant from treated soil (kg/ha-yr),
NR = recharge to groundwater beneath the treated soil (m3/m2-yr, or
m/yr),
OK = concentration of contaminant in water infiltrating through the
treated soil (i.e., leaching from the site) (kg/m3), and
10'000 ~ constant to convert units from (kg/m2-yr) to (kg/ha-yr).
From Appendix A, the concentration of contaminant in leachate is related to the total
concentration (by volume) of contaminant in soil as:
Cfee - Ctl [BD-KD + 0w + H*a]
where:
Ct = total concentration of contaminant in treated soil (kg/m3),
0W = water-filled porosity of soil (dimensionless),
H = Henry's Law Constant for contaminant (dimensionless), and
0» = air-filled porosity of soil (dimensionless).
B-l
-------
This flux of contaminant must be fianslated into a first-order loss coefficient so that:
dCt
_ £ _ p- ft
~77 ~ ~Klee^t
» • at
where:
Kfc,. = first-order loss rate coefficient for leaching (yr1),
Q = total concentration of contaminant in soil (kg/m3), and
t . = time (yr).
Kleo is estimated with the approximation:
K =
which is equivalent to:
K -
to ~
where:
= mass of contaminant in soil (kg),
At = one year,
FA, = flux of contaminant leaching to groundwater (kg/ha-yr),
A = area of land application site (m2), and
10^ = constant to convert units from (m2) to (ha).
JfA-(Vt/d,), this result can be combined with results from Appendix A to yield:
ct di [BD KD •+ ew •+ H-BJ d.
We use this equation to predict contaminant loss to leaching for both land application and the
monofill prototype for surface disposal.
B-2
-------
APPENDIX C
'.. ._ m
Calculation of a "Square Wave" for the Groundwater Pathway
We estimate potential human exposure and risk through the groundwater pathway for both
land application and surface disposal of sewage sludge. To prepare input for the VADOFT
model of contaminant transport through the unsaturated zone, we conservatively assume for both
land application and surface disposal that pollutant is consistently loaded into the top of the
unsaturated zone at the maximum rite estimated by mass balance calculations. We constrain the
duration of this constant pulse, or "square wave" so that the total mass of contaminant leaching
or seeping from the site is conserved. Although our general approach is the same for both land
application and surface disposal, destails differ according to which management practice is being
considered. This appendix provides a brief discussion of our methods for estimating the
magnitude and duration of the "square wave" of pollutant loading for land application and both
prototype facilities for surface disposal.
C.I ., Land Application.
..?:*
Both metals and organic contaminants accumulate in soil with repeated applications of
sludge. We assume that all .competing loss .processes for contaminant in soil can be
approximated as first-order, and that^coefficients describing the rate of loss to each process can
be summed to yield a total or "lumjped" coefficient for first-order loss. Losses at any time t can
then be described as: -:; •
dMt
where:
M. = mass of contaminant in treated soil at tune t (kg) and
= lumped, first-order loss rate for contaminant (yr1).
If contaminant loading to treated soil is approximated as a continuous process, the mass
of contaminant in soil after t years of applications can be described by:
"
Mt = fPA e~Kttf dx = — (1 - e~K' ,
t j v
0 ^tot
where:
PA = total annual loading of contaminant to site (kg/yr).
As t approaches infinity, M, therefore approaches (PA)/^ and yearly loss approaches yearly
loading.
C-l
-------
For organic contaminants, we assume that sludge has been applied repeatedly until
steady-state is achieved. In other words, contaminant has accumulated in the soil until total
yearly losses through runoff, degradation, leaching, and volatilization (which are assumed to be
proportional to the concentration in soil) catch up with yearly loadings of contaminant to soil.
Our estimates of risks from organic contaminants on land application sites are derived for this
steady-state condition. The amplitude of the square wave^pulse for the groundwater pathway
model is therefore equal to the^annual loading of contaminant multiplied by the fraction of
annual loss attributable to leaching. The length of the square wave is equal to the length of the
simulation (300 years). t- ' • % 'v? r-r- - ':i- •"'•*•'"-•"•'•-
For metals, this condition of steady-state is not necessarily reached. According to the
loss coefficients calculated in .this analysis, arsenic is lost most rapidly from treated soil, and
lead least rapidly. With a lumped loss rate of 0.12 per year, arsenic approaches a steady-state
concentration equal to about 8 times its annual loading. After about 10-20 years, yearly losses
closely approximate yearly loadings of 1 kg/ha. Lead, the least mobile of the metals evaluated,
is depleted from treated soil at an estimated annual rate of only 0.0073 per year. If lead were
applied repeatedly to the soil, its expected concentration would increase significantly each year
for the first 500-600 years of repeat applications; yearly losses would not begin to approximate
yearly loadings for several centuries,.
Pollution of groundwater by metals-from sludge depends not only on the cumulative
loading of metals, but also on the period of time in which this cumulative loading takes place.
. We assume that metals are loaded into treated soil through N consecutive, yearly applications
of sludge; after N years, applications are discontinued. To capture the risks associated with the
peak rate at which contaminant leaves the soil layer, we use this peak rate (calculated for the
Nth year of application) for the calculations. We calculate the length of the square wave by
dividing the total (cumulative) loading of contaminant by this maximum rate of loss:
N PA _ N
PA (l-€
where:
TP = duration of "square" wave for approximating the loading .of contaminant
into the top of the unsaturated soil zone (yr).
C.2 Surface Disposal: Monofill [Prototype
Our modeling of the groundwater pathway for the monofill prototype of surface disposal
is similar to that for land application. For both cases, we; assume the site receives repeated
loadings of contaminant for the duration of its active lifetime. By analogy with the above
discussion for land application, this maximum rate of loss from the facility can be described as
a function of its yearly loading, yearly loss, and number of years of active operation:;
C-2
-------
LE ;...== actiye ]$Fetimeof monofffl (yr),
»V *>»-.'^*.: v '" •- v"'-''" ".£,• •-•• . .,; ,f-' ' w. "
"^^I:^^^11^-?11 monofill at end of year LF (kg), and
loaded into the monofill (kg/yr).
of time tMs m^xiniiiin rate of loss c»»ld:.be:mamtained is then:
.C.3 Suface DK|^|: Surface Itinpidundment Prototype
- • • -..w•;.,...-. -..-.i'.-OT®.,~.,~:v*!p?-s-v:-.r's^f^ " of'^%%^>osa^»- our.calculations-atfe'based'::
'•#/ v oti. ^ 'fconsefyaai^e iassuniptidnithlSl^^^tate is mainti^B«! for concentrations of pollutants':
• within the Uquiff.iand sediniient >yerl!?orine -impbuntimeiat. ;We aliso'assume that the-flux of-
cpn^iuit^t.le^iimng from'thl Jmjibu^dnient is constant wiii^espect to time, at least until the'
. %.: to^Wass of contaminant d^^i^^^ For this prdtotype-,
^ ;the length of tjbe square wave used; for execution 6f the .^AlDQFr model is therefore equat'to.
the total-.mass of Contaminant enteimg the impoundmehfeacl:^eari- multiplied by the expected
lifetime of fihe faciHty.and divicl^i by7 the amount lost each "year:
-..,*<• jp± 32x10-* PA TF _ 32x10-*
'•].?$£> ' ; Mf~t . feet
where:
;.,... _ =^ ajnyersipn factor for (sec) to (yr),
PA ,^ =,^. rafe; tlit: contaminant is loaded into the surface
•'
, ..,=.. Active lifetime of^^site (sec), and
= fraction of each year's cqntainHrant loading lost from the site
year to all processes combined (dimensionless).
C-3
-------
|