Statistical analysis of radiation dose derived from ingestion of foods

MISSING IMAGE

Material Information

Title:
Statistical analysis of radiation dose derived from ingestion of foods
Physical Description:
viii, 209 leaves : ill. ; 29 cm.
Language:
English
Creator:
Dougherty, Ward L., 1963-
Publication Date:

Subjects

Subjects / Keywords:
Environmental Engineering Sciences thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Environmental Engineering Sciences -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2001.
Bibliography:
Includes bibliographical references (leaves 206-208).
Statement of Responsibility:
by Ward L. Dougherty.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 026985075
oclc - 47260464
System ID:
AA00018878:00001

Full Text










STATISTICAL ANALYSIS OF RADIATION DOSE
DERIVED FROM INGESTION OF FOODS













By

WARD L. DOUGHERTY


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001





















Copyright 2001

by

Ward L. Dougherty













Dedicated to
Gwen, Justin and Michelle

The Best Family in the World!














ACKNOWLEDGMENTS

I would first and foremost like to thank my wife, Gwendolyn, whose constant

encouragement and support provided the impetus as well as the confidence to finish when

it seemed a daunting task. My best daughter in the world, Michelle, and best son in the

world, Justin, gave me the time and the "I Love You, Dad" at the times I needed them the

most. I see them growing more each day, and this task of writing a dissertation seems

insignificant to my task in parenting. You, as my family, make me more proud than

anything else that I have done in this life.

I would like to offer a special thanks to Dr. W. Emmett Bolch. He provided me

support at the times when I needed it the most and encouragement and support when it

was hard to find.

My professors--Dr. Anghaie, Dr. Wesley Bolch, Dr. Dalton, Dr. Properzio, and

Dr. Lindner--have been invaluable in providing not only their time and assistance but also

their support throughout this endeavor.

My friends--Travis Knight, Gary Chen, Paula Johnson, Katherine Wilson, and

Steve Boddeker--were a great help in providing suggestions as well as listening and

offering insight to help me complete this work.

Much of this work would not have been possible without the assistance of Jerome

Guidry, Brian Birky, and Cindy Hewitt. Their assistance and help have been invaluable

both in the content and the background for this work.














TABLE OF CONTENTS

pane


ACKNOW LEDGM ENTS........................................................................................ iv

ABSTRACT ........................................................................................................... vii

CHAPTERS

1 INTRODUCTION ............................................................................................ 1

Radioactive Dose to the Public......................................................................... 5
Hypothesis........................................................................................................ 5
Goals and Objectives........................................................................................ 6

2 LITERATURE SEARCH ................................................................................. 7

Introduction...................................................................................................... 7
Dietary Intake Data........................................................................................... 7
Concentration of Radionuclides in Food......................................................... 10
Dose Conversion Factors................................................................................ 11
Statistical Considerations................................................................................ 12

3 DIET M ODEL ............................................................................................... 13

Introduction.................................................................................................... 13
Initial M odel................................................................................................... 14
NRC Nuclear Regulatory Guide 1.109............................................................ 14
RESRAD ........................................................................................................ 17
Environm ental Protection Agency.................................................................. 18
United States Department of Agriculture........................................................ 19
Conclusion ..................................................................................................... 21

4 EXPERIMENTATION ................................................................................... 25

Introduction.................................................................................................... 25
Initial Store Samples....................................................................................... 25
Grocery Store Analysis................................................................................... 32
Rice Experimental Analysis............................................................................ 37








H istogram A analysis ........................................................................................ 43
Conclusion and Recommendations for Future Research.................................. 47

5 DOSE CONVERSION FACTORS................................................................. 50

Introduction......................................................................................... ........ 50
Literature Search................................................................................ .......... 50
D discussion ..................................................................... ... ......................... 51
Conclusion and Recommendations ............................................................ 52

6 COMMITTED EFFECTIVE DOSE EQUIVALENT...................................... 54

Introduction ...................................................................................... ...........54
Literature Search ............................................ ......... ............. ......................... 55
Method......................................................................... ................ ............ 55
Analysis of 1990 FIPR Dose Diet............................................................... 57
Grocery Store Data Analysis........................................ ....................... ... 63
Overall Analysis......................................................................................... 67
C onclusions............................................................ ................ ... ................. 72

7 CONCLUSIONS AND RECOMMENDATIONS....................................... 75

APPENDICES

A DATA SHEETS............................................................................. ............ 80

B CRYSTAL BALL OUTPUT DATA .................................... ............... 101

R EFEREN CES ..................................................................................................... 206

BIOGRAPHICAL SKETCH................................................................................. 209













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

STATISTICAL ANALYSIS OF RADIATION DOSE
DERIVED FROM INGESTION OF FOODS

By

Ward L. Dougherty

May 2001

Chairman: Emmett Bolch
Major Department: Environmental Engineering Sciences

This analysis undertook the task of designing and implementing a methodology to

determine an individual's probabilistic radiation dose from ingestion of foods utilizing

Crystal Ball. A dietary intake model was determined by comparing previous existing

models. Two principal radionuclides were considered--Lead-210 (Pb-210) and Radium

226 (Ra-226). Samples from three different local grocery stores--Publix, Winn Dixie, and

Albertsons--were counted on a gamma spectroscopy system with a GeLi detector. The

same food samples were considered as those in the original FIPR database. A statistical

analysis, utilizing the Crystal Ball program, was performed on the data to assess the most

accurate distribution to use for these data. This allowed a determination of a radiation

dose to an individual based on the above information collected.

Based on the analyses performed, radiation dose for grocery store samples was

lower for Radium-226 than FIPR debris analyses, 2.7 vs. 5.91 mrem/yr. Lead-210 had a








higher dose in the grocery store sample than the FIPR debris analyses, 21.4 vs. 518

mrem/yr.

The output radiation dose was higher for all evaluations when an accurate

estimation of distributions for each value was considered. Radium-226 radiation dose for

FIPR and grocery rose to 9.56 and 4.38 mrem/yr. Radiation dose from ingestion of Pb-

210 rose to 34.7 and 854 mrem/yr for FIPR and grocery data, respectively.

Lead-210 was higher than initial doses for many reasons: Different peak

examined, lower edge of detection limit, and minimum detectable concentration was

considered. FIPR did not utilize grocery samples as a control because they calculated

radiation dose that appeared unreasonably high.

Consideration of distributions with the initial values allowed reevaluation of

radiation does and showed a significant difference to original deterministic values. This

work shows the value and importance of considering distributions to ensure that a

person's radiation dose is accurately calculated.

Probabilistic dose methodology was proved to be a more accurate and realistic

method of radiation dose determination. This type of methodology provides a visual

presentation of dose distribution that can be a vital aid in risk methodology.














CHAPTER 1
INTRODUCTION

The purpose of this study was to design and implement a methodology utilizing

the Crystal Ball* program to determine a statistical value, a number, and associated

fluctuation of an individual's radiation dose based on foods bought from local stores in

Gainesville, Florida, and to provide comparison through analysis to a similar previous

study.

The most straightforward approach to an individual's radiation dose

determination has been to use a deterministic approach. The calculation stated below

will provide a committed effective dose equivalent (CEDE) based on ingestion (intake)

of a specific radionuclide, a certain concentration in the food and a dose conversion

factor (DCF).


Intake Concentration DCF = CEDE (Equation 1-1)

where

Intake = individual dietary intake (g/day or g/yr)

Concentration = amount of radionuclide in for (pCi/g or pCi/kg)

DCF = dose conversion factor-term to convert activity in foods ingested
to dose (mrem/pCi).





Crystal Ball is a statistical analyses program written by Decisioneering as an addition to Microsoft
Excel. This provides Monte Carlo sampling of input parameter distribution and trials to determine an
output distribution.








CEDE is a dose quantity that describes the long-term dose to an individual

from an intake of radioactive material (Shleien, Slaback, & Birky, 1998, pp. 3-5).

Depending on what type of dose the individual was trying to calculate, each variable in

the calculation had essentially one, and only one, value determined and set down in the

guidelines by various agencies. The guidelines were put into recommendations, and

these were subsequently, though much later, written into regulatory guidelines. The

advent of better computational methods for radiation dose determination due to better

computer hardware and software and increased amounts of experimental data is now

leading to more accurate determination of the described dose utilizing a probabilistic

approach.

The method of a probabilistic approach versus deterministic approach utilizes

the information that the variables are each described by a statistical distribution. These

statistical distributions are defined by a mean and its associated fluctuations. Figure 1-1

illustrates the methodology and the concept behind this approach. This method provides

a more accurate description of the actual range that each variable might have and the

probability assigned to it.




uS uS a* MMn w : > t i<- .- -< -- _____ n *--

Intake Concentration DCF CEDE Dose
Distribution Distribution Distribution Distribution


Figure 1-1. Probabilistic Method of Dose Calculation








The Health Physics Society has stated in their most recent position papers that

risk assessment must be considered in the context of uncertainties in the estimates

(Burk, 2000, p. 232). This statement is in light of the fact that more people and

organizations are approaching risk-based policy.

This approach allows determination of a final answer, in this case dose, that also

has a distribution. This type of an answer, final dose described by a distribution,

provides a more accurate answer by taking into account the distributions of the

variables with their associated errors and promulgating them through the equation to

come to a final solution.

This document is set up as individual chapters connected by the overall

introduction and conclusion. Each chapter has its own introduction and conclusion.

Additionally, each chapter describes the previous and following chapters to provide a

more unified whole to the reader. The chapters in this document and a brief description

of each are organized as follows:

Chapter 1 (Introduction). This chapter provides an overview of the project of

both its scope and breadth. The types of approaches to radiation dose evaluation are

discussed, both past and present. A brief overview of the different sections of the

chapter is discussed.

Chapter 2 (Literature Search). The applicable literature is cited in this chapter.

The various studies that have been performed that are specific to the radionuclides,

methodology, statistics, and research in this area are reviewed in this chapter.

Additional sources are reviewed to determine current work in the area of food

radioactivity analysis and diet models utilized to determine dose to an individual.








Chapter 3 (Diet Model). The literary references for the individual diet is

provided in this chapter. The methodology for determination of the amended diet is

discussed in this chapter. The diet is described for an individual in this portion of the

paper. A determination of an individual's intake is made in this chapter as well as the

rationale behind the decision.

Chapter 4 (Experimentation). Description of the samples bought, prepared, and

counted from three different local stores is considered in this chapter. Concentration

data for various foods is determined from the experimental data collected. The

distribution of the concentration of radionuclides in food is also undertaken and

resolved in this chapter with the analysis of an additional set of samples.

Chapter 5 (Dose Conversion Factors). Explanation of the Environmental

Protection Agency's Dose Conversion Factors (DCFs) is provided. The discussion of

the various dose conversion factors is considered, and assignment is made to a statistical

distribution to the dose conversion factor variable.

Chapter 6 (Committed Effective Dose Equivalent-CEDE). Analysis of the

previous chapters with regard to the calculation of the CEDE is considered utilizing

Crystal Ball's statistical analysis tools to configure the variables and determine the

output. The various methodologies and analyses on both the original 1990 data and the

newly measured grocery store data are presented to determine the final dose and the

final dose distributions that accompany these data.

Chapter 7 (Results, Conclusions, and Recommendations). The analysis of the

final dose determinations and the program to achieve them are presented.








Recommendations for future work are presented, and conclusions based on the output

from the above chapters are provided.


Radiation Dose to the Public

An individual is expected to get an average annual dose of 360 mrem (Shleieen

1998). Geographic and other factors can change this value from 75 to 5000 mrem.

Individuals are constantly exposed to radiation of all types: cosmic, terrestrial, natural

internal. Without sunlight life itself would be impossible, but radiation has a bad

connotation to the public. People fear the word and the associated images that it

conjures up. The public thinks of Three Mile Island and Chemrnyobi when the issue is

discussed, but radiation is all around us and is a vital and important part of our world.

The plants are the focus of this paper. This dissertation and chapter seek to determine

through theory and experimentation what level of radioactivity is found in our food and

provide a statistical analysis for an improved completeness of description.


Hypothesis

The hypothesis of this dissertation is that the ingestion of radioactivity found in

of foods bought from local stores should be measured and analyzed to determine its

significance. This value that is experimentally measured from foods should have a

statistical value that can be described by a distribution. The final dose that is determined

from these measured values additionally should have a statistical value with a

distribution.









Goals and Objectives

The following provides a list of goals and objectives for this dissertation:

1. Perform a literature search on radioactivity of foods bought in local stores in

Gainesville, Florida, as well as previous studies performed on Florida foods

or the associated radionuclides.

2. Determine a dietary intake of foods based on previous studies.

3. Measure foods bought from three different local stores and determine

radioactivity of lead-210 and radium-226 in these samples.

4. Measure one set of samples to determine distribution to be associated with

concentration of radionuclides in food.

5. Determine the distribution to utilize for the EPA's Dose Conversion Factor.

6. Perform Crystal Ball analysis to determine the final dose, in distribution

form, to the individual from the original FIPR study data and from the

experimentally measured grocery store data.

7. Analyze the results to provide a comparison to the total dose.

8. Compare deterministic and probabilistic methodology of dose calculation.













CHAPTER 2
LITERATURE SEARCH


Introduction

This dissertation covers several fields of study; therefore, a literature search

needs to be performed to find the relevant references in each of these areas. The areas of

search that are involved in this dissertation are dietary intake, food concentration,

radionuclide, dose conversion factors, committed effective dose equivalent, and

statistical distributions. This literature search examines the various literature sources

that were utilized for each of these categories.


Dietary Intake Data

The dietary intake of an individual is quite often information that is specific to

the individual. The dietary intake varies by person due to individuality of the person as

well as local customs and availability of food. In an effort to determine dietary intake of

an individual in Florida, the first source that was researched included previous studies

performed in Florida.

A study of radioactivity in Florida foods and the diet of Floridians necessarily

begins with an assessment of study in the field, both past and present. Some of the most

recent work in Florida that has examined radioactivity in foods in Florida was

performed by the Florida Institute of Phosphate Research (FIPR) (Guidry, Roessler,

Bolch, McClave, Hewitt, & Abel, 1990). This organization was created by the Florida

Legislature in 1978 to conduct supportive research to the development of the state's

7








phosphate resources. This organization has done studies in this field due to its interest in

the environmental aspect of phosphate mining.

There are three sources that provided information both for the dietary model and

the initial concentration of radionuclides in food grown on phosphate and related lands.

The first document is the 1986 FIPR report that provided the initial analysis of radium-

226, lead-210, and polonium-210 in foods grown on phosphate lands (Guidry, Bolch,

Roessler, McClave, & Moon, 1986). The initial diet model that describes dietary intake

was first presented in this document. The method of analysis and the dose evaluation

were described in this book. Simplified analysis of radionuclide concentration in foods

was performed to determine dose to an individual. Three individuals were considered:

control, local, and maximum. A control individual was a reference individual who

consumes "sampled" foods not from mining-related lands. A local individual consumed

10 percent of his "sampled" foods from phosphate lands and 90 percent from control

lands. A maximum individual consumed 100 percent of "sampled" foods in his diet

from phosphate (clay) lands. This individual reflects worst case scenario (Guidry et al.,

1990, pp. 118-119).

The next two documents were associated with this initial document. Brian

Birky's master's thesis referred to the previous document and used the same

methodology, diet, and radionuclides to determine dose attributable to technological

enhancement of this phosphate reclaimed land (Birky, 1990). This document detailed

the previous methods and studies that were utilized to prepare, enclose, and measure the

experimental samples. This thesis discussed the methodology utilized to calculate the








dose to an individual directly from the dietary intake spreadsheet. This thesis was much

more descriptive in the details of diet and dose calculation than the initial paper.

The 1990 FIPR paper was a continuing study based on the recommendations of

the 1986 FIPR paper mentioned above (Guidry et al., 1990). This paper utilized the

same basic dietary intake model of the initial study. Some of the same radionuclides

were considered. This document focused primarily on three radionuclides and five land

types. Three types of individuals were considered in this paper also: local, control, and

maximum. The basic dietary model, with few revisions, was presented in this paper.

This paper analyzed the differences in the data as well as performing regression analysis

on the collected data. This paper had more data points added and more analysis

performed that detailed the soil-to-plant transfer model and refined the dietary intake

model.

FIPR has continued to improve its database with more samples since this report,

and the extended database will be available in a publication in the near future. The

current research and work also has continued to smooth the statistical data.

An important point was brought about by direct discussion with Dr. Birky:

sampled versus nonsampled diets (B. Birky, personal communication, March 13,2001).

A vital consideration in any analysis is the thought given to what foods were and were

not sampled and how to consider them in the final dose determination. These papers and

the subsequent meeting provided invaluable insight into this particular point.

These papers primarily discuss the various analyses performed on foods grown

on Florida lands in general and phosphate lands in Florida in particular. These were the

most useful and pertinent with regard to this analysis.








Concentration of Radionuclides in Food

The concentration of radionuclides in food has been studied in several areas and

contexts. The previous three papers discussed this very subject and determined the

concentration of several radionuclides in various foods grown on phosphate-related

lands.

A study funded by FIPR and performed by the Audubon Society studied the

concentration of radium-226 in alligators, armadillos, and soft and hard shell turtles

(Pritchard & Bloodwell, 1986). These data relate that the hazard from eating these

mammals on mine-impacted lands is unclear.

Dietary intake of lead-210 has been discussed in several articles. Linsalata

(1994) discusses human exposures along plant and animal pathways to thorium,

uranium, radium, lead, and polonium. The exposure pathways were considered, and the

author states that much more work needs to be done in assessing the transfer of lead-

210 and polonium-210 in the human food chain.

Morse and Welford (1971) discusses the dietary intake oflead-210 in the diet of

New York city residents with a result of 1.2 pCi lead-210 per day. This food diet only

included 19 food items. An interesting note is the fact that the concentration of lead-210

was calculated as 0.70 pCi lead-210/kg food.

An analysis was performed on radionuclide contact in Hong Kong food. Yu and

Mao (1999) gives an excellent description of the types of gamma spectroscopy system

used. The diet model and the results were detailed in tabular form with seven

radionuclides under examination. Potassium-40 was found in all solid food and drink








samples. Lead-210 was measured as being greater than half the contribution to the dose

from natural radioactivity.

Carvalho (1995) analyzed the Portuguese population for intake ofpolonium-210

and lead-210. The author primarily examined these two radionuclides and their

ingestion rates for the population. This paper supports the postulation of a different diet

model for a different population. The point is also made in this paper that cooking the

various food prior to eating is not taken into account.


Dose Conversion Factors

The data for the dose conversion factor (DCF) came from four sources. The first

source was Federal Regulatory Guide No. 11 (EPA, 1988, pp. 155-179). This document

provides the methodology used to calculate the DCFs for inhalation, submersion and

ingestion. The tables of the various DCF data for various radionuclides is included in

this manual.

The next two articles, International Council on Radiation Protection (ICRP) 68

(ICRP, 1994) and 72 (ICRP, 1996), provide age-dependent DCFs for workers and

members of the public from intake of radionuclides. Although they were examined for

the purposes of this report, the new DCFs were not utilized for the purposes of

consistency.

The fourth reference for these data was a solution manual that calculated a dose

conversion factor for strontium-90 (Turner, Bogard, Hunt, & Rhea, 1988). This was

utilized as a reference to describe the method to obtain a dose per unit intake factor

from the initial data.








Statistical Considerations

Statistics play a major role in the analyses of this study. Information to utilize

Crystal Ball comes from the manual provided with the program (Decisioneering, 1996).

The manual provides the instruction to utilize the program as well as examples to

familiarize the novice with the operation of the various tools built into the software.

These sources provided the nucleus of the reference material researched to

obtain the necessary data for the background to perform this research and analysis. The

following chapter discusses the diet model and how it was determined. The next chapter

discusses the dose conversion factor to determine which dose conversion factor to use

and what distribution to assign to this value for an accurate estimate of dose

distribution.













CHAPTER 3
DIET MODEL


Introduction

This chapter is a literature search and subsequent analysis of various diet models

currently utilized by regulatory agencies and other organizations. The object of this

portion of the work is to determine the most accurate diet model for calculating radiation

dose to individuals in Florida from their dietary intake.

The choice of a proper diet model is vital to determine dose to an individual from

ingestion. Diet model is not as accurate a term as dietary intake model. This distinction

may seem small, but it is significant. A diet model refers to the assumed or predicted

intake of certain sources of food to individuals. Conversely, a dietary intake model

utilizes surveys of the public to ascertain actual food intake. The goal of this chapter is to

ascertain, using major and well-established sources, the most accurate and comprehensive

source for the dietary intake model for use with our dose evaluation model.

Numerous sources were researched to obtain this goal. The major sources

researched were the Pennington Model, Nuclear Regulatory Commission, the United

States Department of Agriculture CHFSII 1976-1978 Study, the 1994-1996 NFCS study,

RESRAD, and the Environmental Protection Agency. All of these were studied to

determine the most suitable diet model for use with a program to determine dose to the

individual.








Several factors are considered in an effort to determine the right dietary intake

source for our program and dose determination program. The factors that will be

considered to decide on the right source for the final dietary intake model will be the

following: the date of the publication, the sources of the publication, the comprehensive

quality of the data, and the compatibility with the previous FIPR model.


Initial Model

The diet model used to calculate radiation dose in the 1986 and 1990 FIPR studies

was the 1983 Pennington dietary intake model (Pennington, 1983). The Pennington diet

model was derived from the Food and Drug Administration's Total Diet Study. The most

recent revision of this study was based on data from the 1987-88 National Food

Consumption Survey (Pennington, 1992). This information was discussed in the 1990

FIPR study dealing with radioactivity of foods grown on phosphate lands (Guidry et al.,

1990). The dietary intake model used only the adult male category and regrouped the 201

items in the Pennington diet intake model. Table 3-1 shows the data from that paper.

As can be seen from this table, there are 17 major categories and 43

subcategories. It detailed intake in grams per day. The sources that are being studied need

to be examined to determine unit compatibility and possibility of improvement over this

initial model.


NRC Nuclear Regulatory Guide 1.109

The first alternate source examined was the Nuclear Regulatory Commission. The

mission of the U.S. Nuclear Regulatory Commission (NRC) is to ensure adequate

protection of the public health and safety, the common defense and security, and the








Table 3-1: Diet Food Items from Pennington Diet (Guidry et al., 1990)


Source Intake
Item (g/day)

DAIRY
Milk 280.99
Cheese 22.41
MEAT
Beef 129.27
Pork 39.54
Other 69
FISH 20.06
EGGS 30.95
CEREAL FOOD___
Corn Grain 5.18
Grain 4.55
Cereals/Bread 174.7
CAULIFLOWER/BROCCOLI___
Cauliflower 0.71
Broccoli 2.8
LEAFY/COLE VEGETABLE
Cabbage 7.04
Collard Greens 0.45
Lettuce 23.38
Mustard Greens 0.45
Spinach 3.28
Turnip Greens 0.45
Other 0.76
Celery 0.62
LEGUMES____
Green Peas 7.29
Other Beans 25.71
Nuts 4.94
Other 11.28


Source Intake
Item (g/day)

SEEDS/GRAINS
Blackeyed Peas 5.61
Rice 22.94
Yellow Corn 14.41
TUBERS/ROOTS
Carrots 2.92
Onion 4.19
Radish 0.32
Turnip 0.42
Potatoes 85.22
Other 1.1
GARDEN FRUIT
Cucumbers 2.62
Greens Beans 8.8
Green Peppers 1.99
Strawberries 1.23
Tomato 25.18
Watermelon 3.44
Yellow Squash/Zucchini 1.26
Other 6.55
TREE FRUIT____
Citrus
Orange 85.26
Grapefruit 7.78
Lemon 10.71
Other 60.36
SOUPS 36.82
CONDIMENTS 54.12
DESSERTS 78.3
BEVERAGE 1172.44
WATER 512
TOTAL 3071.8








environment in the use of nuclear materials in the United States (NRC, 2000). This board

and its associated bureaucracy accomplish this mission by presiding over the various

aspects of reactor operation, siting, and licensing. Numerous programs and regulations

are employed to determine the safety and feasibility of siting and operating a plant. NRC

Regulatory Guide 1.109, Calculation of Annual Doses to Man from Routine Releases of

Reactor Effluents for the Purpose of Evaluating Compliance with 10 CFR Part 50,

Appendix I, employs a dietary intake model to determine dose to an individual from

ingestion of radionuclides (NRC, 1977). Tables 3-2 and 3-3 list the consumption of

various foods such as fruits, vegetables, meat, and milk.

This information is listed for various groups of people: child, teen, and adult.

Additionally, each table details the intake for the average individual and the maximum

individual. The data in these tables come from ICRP Pub # 23 from 1975 (NRC, 1977)

and AER USDA report of 1974 (NRC, 1977). There are only six categories displayed in

Table 3-2 and seven categories in Table 3-3. The data are limited when compared to the

Pennington dietary intake model. The units are compatible when converted.



Table 3-2: NRC NRG 1.109 Average Individual Intake (NRC, 1977)

Source Child Teenage Adult


Fruits, Veg and Grains (kg/yr) 200 240 190
Milk (L/yr) 170 200 110
Meat and Poultry (kg/yr) 37 59 95
Fish (kg/yr) 2.2 5.2 6.9
Seafood (kg/yr) 0.33 0.75 1
Drinking Water (L/yr) 260 260 360








Table 3-3: NRC NRG 1.109 Maximum Individual Intake (NRC, 1977)

Source Child Teenage Adult


Fruits, Veg and Grains (kg/yr) 520 630 520
Leafy Veg (kg/yr) 26 42 64
Milk (L/yr) 330 400 310
Meat and Poultry (kg/yr) 41 65 110
Fish (kg/yr) 6.9 16 21
Seafood (kg/yr) 1.7 3.8 5
Other Seafood (kg/yr) 1.7 3.8 5
Drinking Water (L/yr) 510 510 730


RESRAD

RESRAD was researched next. This program was designed by the Environmental

Assessment Division of Argonne National laboratories. It was approved by the

Department of Energy for evaluation of radioactively contaminated sites (ANL, 1989).

This code has undergone several benchmarking analyses. The first release of the code

was in 1989. It allows either user input of numerous variables or default variables. Table

3-4 shows these data in tabular form. As can be seen from this table, there are only five

food groups in this diet.


Table 3-4: RESRAD Dietary Intake Parameters (ANL, 1989)

Source Default Units Min Max

Fish 5.4 kg/yr 0 1000
Other seafood 0.9 kg/yr 0 100
Fruit, Veg and Grain 160 kg/yr 0 1000
Leafy Vegetable 14 kg/yr 0 100
Meat and Poultry 63 kg/yr 0 300








There are various sources for this chapter. The seafood data come from the EPA

recommendation of two reports performed in 1981 and 1982. The data for the fruit,

vegetable, and grain come from the EPA 1990 paper which was derived from two earlier

documents: Foods Commonly Eaten by Individuals: Amount Per Day and Per Eating

Occasion and Food Consumption: Households in the United States, Seasons and Year

1977-1978. Additionally, since the 1990 document did not address grain consumption,

these data were taken from the NuReg 1.109 discussed above. For the leafy vegetable

consumption rate, the code refers back to NuReg 1.109 and the average individual. The

meat and poultry consumption rate is also determined from NuReg 1.109 and the EPA's

1990 paper with comparison to another paper by Gilbert et al. from 1983. This database,

like the NRC model, has a limited number of data points. The units are compatible, but

due to the data point limitation, it is not considered as feasible for our diet intake model

(ANL, 1989).


Environmental Protection Agency

The Environmental Protection Agency (EPA) has numerous documents, as is

evidenced by the above discussion of RESRAD information sources. The most

applicable of them is The Exposure Factors Handbook (EPA, 1997). This was first

published in 1989 with its update published in 1997. This paper has a wealth of

information about food consumption by area, age, sex, and various foods; however, this

information is far too voluminous to include in this chapter. The key study utilized to

perform this analysis was the EPA analysis of 1989-1991 USDA CSFII study (USDA,

1996). It should be noted here that this analysis in the exposure factor handbook also








included mean and standard errors that might be useful in statistical analyses performed

as an extension of this work by future researchers.

The EPA's Federal Guidance Report # 13 : Cancer Risk Coefficients for

Environmental Exposure to Radionuclides (EPA, 1995) was also researched but was not

seriously considered due to the fact that the intake considered is not broken down by food

groups, and the units are in kcal/day, units that are not compatible with our comparison.

The limited number of data points also make it incompatible with the previous study.


United States Department of Agriculture

The United States Department of Agriculture (USDA) conducts several food

consumption surveys at regular intervals. One of these was mentioned as a reference in

the EPA Exposure Factors Handbook (EPA, 1997). Numerous articles discuss the various

surveys performed by the USDA (Borrud, Enns, & Mickle, 1996) as well as the trends in

food and nutrient intake that are derived from them (Enns, Goldman, & Cook, 1997).

This last referenced article compares the 1977 NFCS, the 1991 CSFII, and the 1995

CSFII. There are two major USDA food intake survey projects. They are the Nationwide

Food Consumption Survey (NFCS) and the Continuing Survey of Food Intake by

individuals (CFSII). The NFCS is conducted approximately every 10 years with the most

recent performed in 1987-1988. It should be noted here that the original food intake

model in the 1990 FIPR report came from the Pennington model. This model was derived

from two other studies, one of which was the NFCS study.

The most recent CFSII was conducted in 1994 -1996. Over the course of the

three-year study over 16,000 individuals were queried about their dietary intake on two

nonconsecutive days. Obviously, this study produced a large amount of data. These data








were separated in much the same way as the EPA data discussed above. This study has

more data points than that of the EPA, NRC or RESRAD studies, but it only has

essentially 11 major food categories and 15 subcategories. Table 3-5 shows these data for

a one-day sampling of male respondents.


Table 3-5: 1994-1996 CFSII Dietary Intake Data

Source Intake
Item (gm/day)

Total Grain Products 361
Yeast Breads and Rolls 63
Cereals and Pasta 89
Ready to Eat Cereal 16
Mixtures mainly grain 128
Total Vegetables 242
Dark Green Vegetables 14
Deep Yellow Vegetables 8
Tomatoes 37
Total Fruits 172
Citrus Fruits 65
Bananas 19
Non Citrus juices and nectars 19
Total Mlk and milk products 256
Total fluid milk 178
Whole milk 54
Lowfat milk 85
Skim milk 35
Milk Desserts 33
Cheese 18
Total meat, Poultry and Fish 275
Beef 38
Pork 15
Mixtures mainly meat, poulty and Fish 137
Eggs 23
egumes 31
Nuts and Seeds 4








The source of these data, although statistically more accurate in that it came

straight from a survey of a large number of respondents, should also be suspect for the

simple reason that it is a survey. Surveys have their own inaccuracies due to the people

questioned and the method of questioning.


Conclusion

The above databases show that there are a limited number of choices for a

comprehensive source for our diet model. The most promising beside the original diet

model, the Pennington dietary intake model, is the USDA 1994-1996 CFSII database.

This has the largest number of food groups compared to the other databases. The units,

grams per year, are the same as the initial diet model. The Pennington model source was

from the United States Department of Agriculture NCF study. This database comes from

the same organization and the database is newer with more respondents surveyed. These

reasons provide that the final source for the dose estimate program should be either the

original Pennington Model or the diet from the USDA CFSII 1994-1996. A newer

version of the Pennington model would be ideal, but foregoing this possibility the author

chooses to utilize the existing Pennington model minimally updated with data from the

CFSII 1994-1996 survey. These data are shown in Table 3-6. Should a newer more

complete version of the Pennington model become available another comparison will be

performed to determine the most suitable model.

Table 3-7 shows the various factors considered in a decision matrix to allow

determination of a dietary intake model. Five factors were important considerations. The

most important factor was the possibility of similar studies being performed. The initial

Pennington model had the dose calculation performed in 1990. The date of publication








was the next factor. Two models had more recent publication dates but due to the

following three factors were unsuitable.


Table 3-6: Dietary Model Intake


Source Intake
Item (g./day)

DAIRY
Milk 193
Cheese 18
MEAT_____
Beef 37
Pork 15
Other 217
FISH 14
EGGS 24
CEREAL FOOD____
Corn Grain 5.18
Grain 4.55
Cereals/Bread 174.7
CAULIFLOWER/BROCCOLI
Cauliflower 0.71
Broccoli 2.8
LEAFY/COLE VEGETABLE
Cabbage 7.04
Collard Greens 0.45
Lettuce 23.38
Mustard Greens 0.45
Spinach 3.28
Turnip Greens 0.45
Other 0.76
Celery 0.62
LEGUMES
Green Peas 7.29
Other Beans 25.71
Nuts 4.94
Other 11.28


Source Intake
Item (g./day)

SEEDS/GRAINS
Blackeyed Peas 5.61
Rice 22.94
Yellow Corn 14.41
TUBERS/ROOTS ____
Carrots 2.92
Onion 4.19
Radish 0.32
Turnip 0.42
Potatoes 85.22
Other 1.1
GARDEN FRUIT ____
Cucumbers 2.62
Greens Beans 8.8
Green Peppers 1.99
Strawberries 1.23
Tomato 25.18
Watermelon 3.44
Yellow Squash/Zucchini 1.26
Other 6.55
TREE FRUIT
Citrus
Orange 85.26
Grapefruit 7.78
Lemon 10.71
Other 60.36
SOUPS 36.82
CONDIMENTS 54.12
DESSERTS 78.3
BEVERAGE 1172.44
WATER 512
rOTAL 2997.58








Table 3-7: Decision Matrix Table

Model Similar Date of Source Quality Comparability
Studies Publication_______________
Pennington 1990 Yes 1990 1987/8 17 and 43 Good
NRCNRG 1.109 No 1977 1974 5 or 7 Fair
RESRAD No 1989 1977 5 Fair
EPA-EFH No 1997 1989/91 .-- Poor
EPAFRG # 13 No 1995 1974 -- Not


The source of each model was examined as the third factor. The EPA Exposure

Factor Handbook had the most recent source with the Pennington model second.

Quantity, or number of food categories for overall diet and possibility to analyze them,

was considered with Pennington having the most complete diet and most available food

categories. Both EPA documents in this category were not applicable due to the large

number of categories and dietary items. Comparability was considered as the fifth and

lowest priority category. This column relates to the format in which the data are

presented and similarity of individual studies. As an example, EPA's FGR #13 is in Kcal

per day, which is difficult to compare with grams or kilograms per day. The Pennington

had good comparability, whiel NRC 1.109 resrad were listed as fair due to having

comparable units without a specified individual such as a 25-year-old male. All of this

led the author to a determination of an updated Pennington diet model as the best choice

for the dietary intake model.

A dietary intake model choice is made more difficult due to the fact that it is hard,

if not impossible, to define a "normal" individual or diet. This is even more complicated

when a limited area such as Florida or Gainesville is chosen. The closest similar previous

study to determine individual doses in Florida was the Pennington model. Therefore, an

updated Pennington model was chosen.








The errors associated with the various diets were not included only the EPA EFH

had errors associated with the dietary intake. The analyses performed in the following

chapters assign various distributions which include fluctuations, errors, and variability.

The distribution for this factor is assumed to be a lognormal due to the facts that

dietary intake is a variable in which the individual has a wide latitude of intake and

therefore some will exercise this power. Allowing for this fact, the factor was evaluated

as a lognormal distribution as well as a normal (gaussian) distribution.

The next chapter discusses the results derived from the gamma spectroscopy

analysis of the local samples bought and analyzed from the various stores in the

Gainesville, Florida, area.













CHAPTER 4
EXPERIMENTATION


Introduction

This is the fourth in a series of seven chapters the overall goal of which is to

design and implement a methodology utilizing the Crystal Ball program to determine a

statistical value, a number, and associated fluctuation of an individual's radiation dose

based on foods bought from local stores in Gainesville, Florida, and to provide

comparison through analysis to a similar previous study. This dose will be described not

by a singular number but will be expressed as a distribution. This distribution will be

determined by a series of analyses, on both the 1990 FIPR diet and a new set of data

determined by experimentation utilizing the Crystal Ball program to evaluate the

distribution and value of the final dose.

Crystal Ball is a forecasting program that is an "add-on" program to Microsoft

Excel (Decisioneering, 1996). Initially designed as a financial forecasting program for

business analysts, this program has a unique and powerful ability to determine the dose

distributions that are under investigation. A Monte Carlo random sampling technique is

utilized within the program to determine the distribution of the final outcome.

The goal of this chapter is to describe the analysis performed on various samples

to determine the concentration and distribution of radionuclides in the foods purchased.

This goal will be accomplished in several stages. Previous work and literature search was

undertaken to determine what to measure, how to measure it, and what foods to analyze.








Next, a discussion of the various radionuclides under consideration will be reviewed.

Then, the actual experimental analyses will be described. Four separate analyses were

performed to measure the concentration of radionuclide concentration in food. The four

analyses will be described along with the information obtained. The conclusion will

consolidate all the data


Initial Store Samples

Literature Search

The literature search described in the second chapter of this series described the

previous samples taken from various locations to assess the radionuclide concentrations

in foods grown in several regions around the world. The focus of this chapter was

necessarily limited to a study of radionuclide concentration of foods in Florida. The

additional sources of literature provided useful comparisons on the various radionuclides

considered, the diets studied, and the methods of analyses.

There are three sources that provided information both for the dietary model and

the initial concentration of radionuclides in food grown on phosphate and related lands.

The first document is the 1986 FIPR report that provided the initial analysis of radium-

226, lead-210, and polonium-210 in foods grown on phosphate lands (Guidry et al.,

1986). Other radionuclides were also examined in these data. The diet model that

describes dietary intake was first presented in this document. The method of analysis and

the dose evaluation were described in this book. Simplified analysis of radionuclide

concentration in foods was performed to determine dose to an individual. Three types of

individuals were considered: control, local, and maximum.








The next two documents were associated with this initial document. Brian Birky's

master's thesis referred to the previous document and used the same methodology, diet,

and radionuclides to determine dose attributable to technological enhancement of this

phosphate-reclaimed land (Birky, 1990). Birky (1990) detailed the previous methods and

studies that were utilized to prepare, enclose, and measure the experimental samples.

Additionally, the methodology utilized to calculate the dose to an individual

directly from the dietary intake spreadsheet was explained. The explanation was much

more descriptive in the details of diet and dose calculation than the initial chapter.

Guidry et al. (1990) conducted a continuing study based on the recommendations

of the 1986 FIPR paper. The same basic dietary intake model was used as that considered

in Guidry et al. (1986). The same radionuclides were considered. Three radionuclides and

five land types were examined. Three types of individuals were considered in this paper

also: local, control, and maximum. The basic dietary model, with few revisions, was

presented in this paper. The differences in the data were analyzed. Regression analysis

was performed on the collected data. More data points were added from the previous

study, and more analyses were performed that detailed the soil to plant transfer model

and refined the dietary intake model.

FIPR has continued to improve its database with more samples since this report,

and the extended database will be available in a publication in the near future. The

current research and work also has continued to smooth the statistical data. These

chapters primarily discuss the various analyses performed on foods grown on Florida

lands in general and phosphate lands in Florida in particular and are the most useful and

pertinent to this study.








The concentration of radionuclides in food has been studied in several areas and

contexts. The previous three chapters discussed this very subject and determined the

concentration of several radionuclides in various foods grown on phosphate-related lands.

To provide for data for this report, and as a means of comparison for the previous

report, similar samples were taken from local stores. This chapter will discuss the

radionuclide considerations, the stores utilized, the samples taken, the method of

measurement, and the results of the measurements.


Radionuclide Considerations

Three radionuclides were considered consistently in the previous reports: lead-

210, radium-226, and polonium-210. Each of these is of concern for various reasons. All

are from the uranium decay chain, and two are progeny of radium-226. The lead-210 and

radium-226 radionuclides will be discussed in turn.

Lead-210 (Pb-210)

This radionuclide is a progeny of radium-226 through decay of radon-222. Lead-210,

unlike radon, is a reactive radioisotope that adsorb onto particulates and therefore pose a

possible risk to humans through ingestion and inhalation. Most environmental lead is

associated with sediments, and the rest is in dissolved form. Short-term exposure to even

low levels can cause changes in red blood cell chemistry; developmental problems; and

attention span, hearing and hearing and learning disabilities in children. Adult short-term

exposure can cause a slight increase in blood pressure. Long-term exposure has been

linked to cerebrovascular and kidney disease (Weiner, 2000, pp. 221,222).

Environmental and toxic considerations aside, a large fraction of the lead-210 in the

environment have been formed following the decay of radon-222. Therefore, higher








concentrations of lead-210 are found in the surface soils. This increases the chance of

intake through the human food chain adding to an individual's dose (Harley, 1988).

Additionally, lead was analyzed in the previous FIPR studies and provides a point of

comparison for the experimental data obtained in this study.

Radium-226 (Ra-226)

There are more data on this radionuclide than on any other radionuclide.

Inhalation of radon daughters account for 55% of the human exposure to natural sources

of radiation (Shleien et al., 1998). Radium toxicity is related to bone sarcomas and sinus

sarcomas due to its competition for bone with calcium. These factors as well as the fact

that the FIPR database includes this radionuclide led to the consideration of radium-226

as one of the points for analysis in this study.


Original versus New Database

There are two databases that could have been considered from the FIPR studies of

the previous radionuclides. The original database from the 1990 FIPR report was chosen

because the newer database has not been completed, confirmed, or published. The

original database considered all three radionuclides and the diet model and has been in

the literature numerous years. These reasons led to inclusion and comparison of the

original database.


Samples Considered

The March 1986 FIPR report analyzed over 100 food samples, replicated up to three

times, collected from 62 land parcels. The Phase 2 1990 FIPR report initial report

collected and evaluated approximately 70 samples from five land parcels. These samples








were considered to determine the samples to evaluate from the stores. The samples are

listed in Table 4-1.


Table 4-1: Food Samples Analyzed from Local Grocery Stores


Beef 1
Beef Kidney I
Black-Eyed Peas 3
Brazil Nuts 1
Brazil Nuts Shells 1
Broccoli 3
Cabbage 3
Carrots 3
Cauliflower 3
Collard Greens 3
Corn 3
Cucumber 3
Eggplant 3
Grapefruit 3
Green Beans 3
Greens Onions 3
Green Peppers 3
rish Creamer Potatoes 1
Lemons 3
Lettuce 3
Lima Beans 3
Mustard Greens 1
Okra 3


onions 3
)ranges 3
Parsley 3
Peas 3
Pole Beans 2
Potatoes 3
Purple Hull Peas 2
Radishes 3
Red Potatoes 2
Rice 3
Spinach 3
Strawberries 3
Swiss Chard 1
Tangerine 2
Tomatoes 3
Turnip Greens 3
Turnip Roots and Greens 1
urnumip Roots 2
Watermelon 2
Yellow Corn 3
Yellow Squash 3
Zucchini 3
Total 113


The samples taken ranged from beef to zucchini. There were 113 samples total;

45 foods were sampled. A sample is considered as a 0.5 marinela beaker filled with the

food in question. Eight foods had only one sample; 6 foods had two samples; and the

remainder of the foods, 31, had three samples. This provided a good average for each

food from the three stores.

Samples were bought from three stores in the local area to provide a better

statistical analysis. Samples were purchased from large supermarkets to increase the








usefulness of this analysis. People outside the Gainesville area and the state of Florida

could utilize these same data in other areas of the country. The first set of samples was

purchased from Publix at 5200 NW 43rd Street on 12 August 2000. Albertsons at 3930

SW Archer Road was the site where the second set of samples was purchased on 13

October 2000. The third store was Winn Dixie at 7303 NW 4th Boulevard where the third

set of samples was purchased on 20 December 2000. It is important to note that some of

the samples only had one or two replicates. This was usually due to the limited

availability of samples due to their seasonality.


Location of Samples

The samples were purchased from the various stores listed above. The question

that should be considered is where they were grown. This is an important factor due to

soil contamination, plant uptake, and therefore plant contamination. Samples from each

individual store come from numerous samples, which often change daily (Greg Sciullo,

personal communication, 30 October 2000). Even if a purchaser asks on the day the food

is bought, the store can usually only provide the supplier and region and not the location

at which the food was grown. This is why it is important to do this and follow-up studies

that consider exact sources and their soil radioactivity as well as plant uptake and human

consumption availability.


Brazil Nuts

Brazil nuts and Brazil nut shells were actually the first product bought sealed and

studied. Of all the samples examined, they had the most number of peaks although not all

were identifiable.








Food Preparation

All foods were prepared as for normal human consumption. No foods were

cooked, and food was cleaned, cut, and sliced to fill individual containers to maximize

weight. The foods were then fit into a 0.5-liter Marinelli beaker. The beaker was capped,

sealed, and stored for two weeks to allow ingrowth of radon-222 and its daughter

products to equilibrium with its parent radium-226. The sample was then weighed and

counted on one of two high-resolution gamma ray spectrometers. The scale that samples

were weighed on was a Mettler P2000N, Serial No. 394916. Detectors 2 and 4 were used

for the analysis. Detector 2 is a Germanium well detector, Serial No. 22P63XC,

University Property No. 491044 1100485. Also, a Germanium well detector, Serial No.

12841211302, University Property No. 4910 AA 117706, is the University Property No.

for Detector 4. The count time varied from 9 to 24 hours. Most samples were counted for

9.5 hours. Sampled items were counted on only two of four detectors available. This was

due to consistency of only using two detectors as well as the limited availability of the

other detectors.


Grocery Store Analysis

After gamma counting the samples utilizing detectors 2 and 4 in the Environ-

mental Engineering Sciences laboratory, a peak search was performed. Each spectrum

was visually inspected for additional peaks that the library search did not recognize.


Radionuclides Evaluated

The 1990 FIPR study evaluated their samples for three radionuclides. This

analysis considered the same three radionuclides because they are from the U-decay








series and have identifiable peaks when counted on a gamma spectroscopy system. These

are associated with phosphate mining and are exposed to the surface and therefore may

be taken up by plants (Guidry et al., 1986). Radium-226 decays through several short-

lived isotopes to radon-222. Radon is a gas that accumulates in structures and can provide

a significant contribution to an individual's dose.


Radium-226 (Ra-226)

The radium content was calculated by summing the three peaks at 295.2, 352.0,

and 609.4 keV. These peaks are from the Pb-210 and Bi-214 daughters. The results are

reported in pCi/gm of material measured (pCi/g)


Lead-210 (Pb-210)

The lead-210 content was calculated utilizing the 10.8 keV peak activity. The

results are reported in pCi/gm of material measured (pCi/g).


Potassium-40 (K-40)

The potassium-40 (K-40) radionuclide was measured, and data are available for

analysis but are not reported in this chapter due to the fact that they were not analyzed in

the FIPR 1986 or 1990 report and therefore have little use in a comparison methodology.

It is interesting to note that the 1460keV potassium-40 peak was present and easily

identifiable in a majority of the samples.


Correction Factor

A correction factor based on the detector, the radionuclide, and the 4513 standard

was calculated. The standard has an activity of 33200 pCi. The correction factor was








determined for radium-226 by combining the counts from the three peaks: lead-214 (295

KeV), lead-214 (352 KeV), and the Bi-214 (609 KeV) and dividing this sum by the time

to obtain the rate of the sample in counts per second (cps). The equation for the

calibration factor is shown below:


Calibration factor (CF) =4513 Activity/Measured count rate (Equation 4-1)


The calibration factors were calculated and are shown in Table 4-2.


Table 4-2: Compilation of Calibration Factors


Pb-210 Ra-226
__________(PCi/cps) (Pci/cps)
Detector 2 3.40E+05 1670
Detector 4 1.10 E+05 1664


Calculation

Once the counts for each sample were analyzed and the blank sample counts were

subtracted, the number was multiplied by the correction factor and divided by the weight

to obtain the answer in pCi/gram. The calculation is shown below:


Concentration = (Sample (cps)-Background (cps))*CF/Weight (Equation 4-2)


Minimum Detectable Activity

Many of the samples returned values of zero at several of the peaks examined.

The minimum detectable activity was calculated at each of these data points and reported

as the actual activity. This methodology provides for a conservative dose analysis as well

as providing a more complete analysis.









A calculation of minimum detectable concentration (MDC) is calculated by

first calculating the limit of detection (LD), as shown below in Equation 4-3.


Limit of Detection (LD) = 2.83 ((Blank/(time))1/2


The MDC is then calculated using the Equation 4-4.


MDC (pCi/gm) = LD (cps) CF (pCi/cps)/Weight (g)


(Equation 4-3)


(Equation 4-4)


Once these data were calculated, they were reported in the data as the counts.


Output of Analysis

Figure 4-1 shows a sample output from the Gamma Vision Program utilized to

count the various samples. An output report similar to this one was produced for each

individual sample. A spectrum was also printed out to allow a visual observation and

comparison with other samples.



Detector #4 ACQ 09-Sep-00 at 9:05:03 RT = 34200.0 LT = 34187.1
Detector # 4 HPGe End Cap in Green Shield
Beef 09/09/2000
ROI# RANGE( keV) GROSS NET CENTROID FWHM FW(1/10) LIBRARY ( keV) Bq
1 72.11 79.55 1190 297 71 76.91 0.59 1.35 No close library match.
2 256.31 261.13 354 -25 34 258.66 0.39 0.71 No close library match.
3 522.34 526.95 457 221 30 524.76 0.66 2.91 No close library match.
4 623.75 626.36 145 56 13 624.77 1.22 1.74 No close library match.
5 1492.05 1497.04 959 768 37 1494.45 2.11 3.73 No close library match.


Figure 4-1: Sample Report for Beef


Results and Analysis

Appendix A illustrates, in tabular form, the raw data that provided the dose

determination. Peak information was determined from reports similar to Figure 4-1.

These were output from the library search performed on each spectrum measured from









each sample. Appendix B contains the Crystal Ball analysis charts for the various

analyses performed.

Table 4-3 lists the results for this analysis. As can be observed from these data,

cucumbers show the highest lead-210 concentration at 47 pCi/g. Beef was observed to

have the lowest lead-210 concentration at 0.076 pCi/g. Potatoes, rice, beef kidney, and

watermelon show the lowest radium-226 concentration with 0.002 pCi/g. Parsley had the

highest observed concentration at 0.029 pCi/g.


Table 4-3: Gamma Spectroscopy Analysis of Local Grocery Samples


Averages (pCi/g)
Item Pb-210 Ra-226
Beef 0.076 0.004
Beef Kidney 0.357 0.002
Black-Eyed Peas 0.494 0.003
Brazil Shells 1.021 0.003
Broccoli 0.829 0.005
Cabbage 0.858 0.006
Carrots 0.555 0.004
Cauliflower 1.466 0.005
pollard Greens 1.372 0.006
Corn 0.506 0.006
Cucumber 47.082 0.009
Eggplant 1.425 0.006
grapefruit 0.096 0.005
Green Beans 3.837 0.007
reen Onions 0.852 0.014
reen Peppers 0.431 0.004

I ish Creamer
Potatoes 0.120 0.006
Lemons 0.414 0.005
Lettuce 1.237 0.004
Lima Beans 0.963 0.003
Mustard Greens 1.622 0.017
Okra_______ 1.072 0.003
Onions ______0.599 0.004


Averages (pCi/g)
Item Pb-210 Ra-226
Oranges 0.682 0.003
Parsley 7.378 0.029
Peas 0.821 0.003
Pole Beans 0.704 0.004
Potatoes 0.846 0.002
Purple Hull Peas 0.454 0.003
Radishes 0.580 0.004
Red Potatoes 0.215 0.003
Rice 0.654 0.002
Spinach 0.422 0.020
Strawberries 0.715 0.004
Swiss Chard 2.010 0.006
Tangerine 0.686 0.004
Tomatoes 1.011 0.003
Turnip Greens 0.941 0.006
urnip Root and Green 1.841 0.005


urnumip Roots 0.941 0.006
Watermelon -0.776 0.002
Yellow Corn .0.680 0.005
Yellow Squash 0.613 0.004
Zucchini .0.108 0.005
Brazil Nuts 1.324 0.007








The average for lead-210 for all samples is 2.037 pCi/g. This value is higher than

most due to the largest value increasing the value. All of the samples have concentrations

below this value with the exception of parsley and cucumber. The average for Ra-226 is

0.006 pCi/g. Fifteen sample concentrations lie on or above this value with all others

measured below. Brazil nuts and their shells cause this value to be higher than most of

the measured values. The concentration of mustard greens lies at this value, and all other

concentrations lie below this value.


Rice Experimental Analysis

Determination of the distribution associated with the concentration of

radionuclides in food was undertaken with the following results. The distribution of a

radionuclide concentration in a food was approached utilizing two major methods:

experimentally and with a review of relevant literature current studies.


Review of Literature and Current Studies

The 1986 FIPR report initially assumed a lognormal distribution. The subsequent

pilot study and analyses utilizing a residual test bore out this hypothesis. This was

performed primarily for radium-226 in the various food items for the study. The study

analyzed 31 food items in six general categories.

The subsequent 1990 study and the current research agree with the initial analysis

from the 1986 paper. The findings were for a lognormal distribution, which was

determined by analyses of the data distribution for specific food items grown on each

land analyzed for each radionuclide.








Experimental Analysis

Even though these data seem conclusive in one method to assume a lognormal

distribution, they do not address the point of different food items, especially the grocery

items. This chapter deals with research performed on grocery store items, whereas the

previous studies did not. The 1990 FIPR report (Guidry et al., 1990) analyzed one sample

of each food type but could not perform any relevant or applicable analyses on only that

one data point. This chapter dealt with 3 data points (in most cases) from each food type.

An analysis was undertaken to analyze a food type for the distribution. The results

obtained, although numerical in nature, were essentially qualitative. The determination of

distribution was obtained through measurement and analysis of rice obtained from Publix

and measured for a single radionuclide.


Choice of Food Sample

The rice was purchased from Publix on 4 February 2001. The brand chosen was

Publix' own brand. This was chosen both for the price consideration as well as the

probability of consumption due to the limited cost by an average consumer. Rice was

chosen also because it provided a consistently high potassium-40 peak on prior analyses.

This provided a good indication that this sample would provide a good consistent peak to

analyze on all samples measured.


Preparation

Similar to all the experimental analysis for this research, the 20 samples of rice

were placed in Marinelli beakers. Each 0.5 beaker was filled with rice, sealed and

refrigerated for two weeks to allow for ingrowths of the Radon daughters to equilibrium.








Measurement

The samples were weighed after the two-week period. Each sample was then

measured on one of two high purity GeLi detectors. Detectors 2 and 4 were chosen for

this analysis due to their previous utilization with other samples as well as their

availability. The samples were each measured for 9-1/2 hours. Once again, this time was

chosen to provide an opportunity for future comparison with other data obtained in this

research.


Comparison

The spectrums of the samples, once counted, were individually examined to

ensure the potassium-40 peak was observed and analyzed. As expected, the peak was

found on every sample to varying degrees. An analysis of the data was performed on the

10 samples measured on each detector as well as overall for all 20 samples.


Results

These data were analyzed for several statistical attributes, such as skewness,

maximum, minimum, kurtosis, range, and standard deviation. Tables 4-4, 4-5, and 4-6

show the summary statistics for detector 2 and detector 4. A comparison of the raw data

and these numbers illustrate that one point on detector 4 was an outlier. Tables 4-7 and 4-

8 show the same statistical comparisons for the combined data with and without the

outlier, respectively.

A survey of the statistics from detector 2 (Table 4-4) shows that the range of

counts per second per gram for the samples measured was 3.36 10-6. The mean of all

samples on these detectors was 5.83 10-6. The standard deviation was 1 10-6. The








skewness, a measure of the distribution to deviate from a standard distribution, is 0.378.

This supports the contention of the literature cited above that states that the distribution of

radionuclide concentration of food is lognormal, in this case positively skewed.


Table 4-4: Summary of Rice Sample Statistics from Detector 2

SUMMARY STATISTICS DETECTOR 2


Mean
Standard Error
Median
Standard Deviation
Sample Variance
Kurtosis
Skewness
Range
Minimum
Maximum
Sum
Count
Largest (1)
Smallest (1)
Confidence Level (95.0%)


5.82591E-06
3.18389E-07
5.71661E-06
1.00684E-06
1.01372E-12
-0.165870891
0.378137914
3.36092E-06
4.26156E-06
7.62247E-06
5.82591E-05
10
7.62247E-06
4.26156E-06
7.20248E-07


Table 4-5 shows similar statistics for the samples measured on detector 4. The

outlier is included in this analysis for the sake of comparison and to illustrate the effect

that the outlier has on the analysis. The mean of the data with the outlier is 5.05* 10-6. The

standard deviation is 1.6 *10"6. The range is 5.96* 10-6. This number is close to twice as

large as the range associated with detector two samples. The mean was measured as

5.05 *10-6 and the skewness as -1.25*10-6. This is not only larger but in the opposite








direction to the distributions both researched and assumed. An observation of the raw

data illustrated that one data point, 1.3 10-6 was an outlier. Once this was removed, Table

4-6 was obtained and analyzed.


Table 4-5: Summary of Rice Sample Statistics from Detector 4

SUMMARY STATISTICS OF RICE SAMPLES ON DETECTOR 4


Mean
Standard Error
Median
Standard Deviation
Sample Variance
Kurtosis
Skewness
Range
Minimum
Maximum
Sum
Count
Largest (1)
Smallest (1)
Confidence Level (95.0%)


5.05372E-06
5.16785E-07
5.21956E-06
1.63422E-06
2.67067E-12
2.670178378
-1.254131791
5.96386E-06
1.29724E-06
7.2611E-06
5.05372E-05
10
7.2611E-06
1.29724E-06
1.16905E-06


Table 4-6 illustrates that with the outlier removed, the mean is now higher at

5.5 106. The standard deviation has now been reduced to 1.02* 10-6. The range is now

3.25 *10.6, less than detector 2. The skewness measured a +0.37, which is very close to

that measured for detector 2. These data, with the outlier removed, tend toward a

lognormal distribution due to the skewness measured and the similarity of data obtained

from both sets of samples.








Table 4-6: Summary of Rice Sample Statistics without Outlier from Detector 4

SUMMARY STATISTICS ON DET 4 WITHOUT OUTLIER

Mean 5.47111E-06
Standard Error 3.40689E-07
Median 5.52044E-06
Standard Deviation 1.02207E-06
Sample Variance 1.04462E-12
Kurtosis -0.449095314
Skewness 0.37326003
Range 3.25132E-06
Minimum 4.00978E-06
Maximum 7.2611 E-06
Sum 4.924E-05
Count 9
Largest (1) 7.2611 E-06
Smallest (1) 4.00978E-06
Confidence Level (95.0%) 7.85631E-07


An analysis was undertaken to compare the same statistics on both data sets

combined. There were two analyses performed, one with the outlier and one without the

outlier. These are shown in Table 4-7 and 4-8.

Table 4-7 illustrates the summary statistics on the rice samples measured from

both detectors combined. Consideration of the outlier would provide a mean of 5.4* 10-6.

A standard deviation would be obtained that would be 1.38* 10-6. The range would be

6.3 10-6 with the minimum and maximum measured at 1.3 10-6 and 7.6* 10-6,

respectively. The skewness would be a 1.2* 10"6. All of this, as well as the data from

detector 4 above, details a reasonable justification for removing the outlier and

considering the other 19 data points as the total sample.








Table 4-7: Summary of Rice Sample Statistics: Total Rice Samples

STATISTICS-ALL DATA POINTS

Mean 5.43981E-06
Standard Error 3.08395E-07
Median 5.49973E-06
Standard Deviation 1.37918E-06
Sample Variance 1.90215E-12
Kurtosis 3.328489995
Skewness -1.201595498
Range 6.32524E-06
Minimum 1.29724E-06
Maximum 7.62247E-06
Sum 0.000108796
Count 20
Largest (1) 7.62247E-06
Smallest (1) 1.29724E-06
Confidence Level (95.0%) 6.45478E-07



Utilizing the other 19 data points and performing a summary analysis, Table 4-8

is obtained. As can be observed from Table 4-8, similar to the observations from Table 4-

6, the mean has now increased to 5.7* 10-6. The standard deviation has now decreased to

1.0"* 10-6. The range, 3.6*10-6 is essentially half of that obtained with the outlier in Table

4-7, and the skewness was increased to a positive 0.32* 10-6, once again supporting the

data obtained by the two sources in the literature search from analyses of previous data.


Histogram Analysis

A histogram analysis was performed on the various sets of data prior to the

summary statistics that were derived above. These initial analyses were performed to

provide a visual representation of the data and their subsequent distributions. They have

been included here to provide for a more complete data analysis as well as to show how

the outlier was initially found. Figure 4-2 shows the data plotted in a histogram. The








outlier lies to the left-hand side of the plot, conspicuously alone. Further consideration

required determination of which detector and sample this number was obtained.


Table 4-8: Summary of Rice Sample Statistics: Total Rice Samples without Outlier

COMBINED STATS EXCLUDING OUTLIER

Mean 5.65784E-06
Standard Error 2.29904E-07
Median 5.52044E-06
Standard Deviation 1.00213E-06
Sample Variance 1.00426E-12
Kurtosis -0.512673412
Skewness 0.318703287
Range 3.61269E-06
Minimum 4.00978E-06
Maximum 7.62247E-06
Sum 0.000107499
Count 19
Largest (1) 7.62247E-06
Smallest (1) 4.00978E-06
Confidence Level (95.0%) 4.83011 E-07


Figures 4-3 and 4-4 show Detector 2 and 4 rice sample histograms, respectively.

Figure 4-3 shows all data in a close grouping and range, whereas the histogram of

detector 4 has a similar grouping with the notable exception of one data point.

The detector 4 histogram has a larger range than that presented on the histogram

of detector 2. The grouping of detector 2 has a much smaller range when the outlier is

excluded. This was the first indication that there was a point that should be removed in

the consideration of the data. Removing this data point and replotting detector 4 rice

samples in a histogram reveals Figure 4-5.









This histogram reveals the reduced range similar to that obtained with the samples

from detector 2. The grouping could be lognormal or normal. A histogram of all samples

as well as a statistical summary analyses performed above is a more accurate indicator of

distribution. Figure 4-6 shows the histogram of the entire data set with the exception of

the outlier.





Histogram of Al Rice Samples

4.5--- --- -----------.--------
4
o 3.5
C 3-
S2.5-
2-
1.5

0.5
0 [] . . .


'V ~ ~ ~ ( -bb ~ .Oj ~
C$3r (P (b ( :b^ C ^ <

Counts/SecondlGram


Figure 4-2: Histogram of All Rice Samples


Histogram Detector 2
.5 .. . . ..... ... . . . . . . . . . . . . . .......__... ..__ ....... ...._ . _ . . . . .

3
o .5
02
15
0.5 -


Coun0econd/Gm

CounisiecondIGrmm


Figure 4-3: Histogram of Detector 2 Samples






























Figure 4-4: Histogram of Detector 4 Samples


Histogram Detector 4 excluding outler

3 .5 ... . ... ........ ... .. ..... ...... .. ......... ...
3.------------------- ---------- ---------------------......-------------------....----------------....................
3
2.5
S 2-
Cr 1.5-
u. 1
0.5
0




Counts/SecondGram


Figure 4-5: Histogram of Rice Samples on Detector 4 excluding Outlier.


This curve illustrates a bimodal distribution around 6* 10-6 counts/gram/sec and

has a slightly lognormal appearance. The distribution at this point was determined using

the summary statistical analysis to ensure that the outlier removal was justified and to








determine the shape of the distribution. Additionally, the data were analyzed using the

Crystal Ball program to determine that the distribution, though limited in number of data

points, most closely approximates a lognormal distribution.




Histogram of Al Rice Samples

4 .5 _..- .. .- ....._ ...-.- .... ----- ----- ----------------- ------ ------- .. ..........




S- 3




Counts/SecondlGram


Figure 4-6: Histogram of All Samples excluding Outlier


Conclusion and Recommendations for Future Research

Analyses were performed on 20 rice samples obtained from Publix Supermarket

in Gainesville, Florida. The 20 samples were tested on detector 2 and detector 4 of the

gamma spectroscopy laboratory at the environmental Engineering Sciences Department

at the University of Florida. The goal of this portion of the research was to determine

qualitatively what type of distribution is exhibited by concentration of radionuclides in

foods. A literature search was performed providing information specific to foods grown

on phosphate related lands. These data stated that the distribution followed a lognormal

distribution. The analyses on these samples, with histograms and summary statistics,

proved that one of the data points was an outlier and, when excluded, provided good
0 2.5-
N. 1.5-
0.5-
0-



CountsWSecondlGram


Figure 4-6: Histogram of All Samples excluding Outlier


Conclusion and Recommendations for Future Research

Analyses were performed on 20 rice samples obtained from Publix Supermarket

in Gainesville, Florida. The 20 samples were tested on detector 2 and detector 4 of the

gamma spectroscopy laboratory at the environmental Engineering Sciences Department

at the University of Florida. The goal of this portion of the research was to determine

qualitatively what type of distribution is exhibited by concentration of radionuclides in

foods. A literature search was performed providing information specific to foods grown

on phosphate related lands. These data stated that the distribution followed a lognormal

distribution. The analyses on these samples, with histograms and summary statistics,

proved that one of the data points was an outlier and, when excluded, provided good








agreement with the data obtained from the previous research. The concentration of

radionuclide in food studied in this set of analyses followed a lognormal distribution.

Further research should be undertaken, both in the form of a literature search and

experimentally, to determine if this is accurate for a wider range of foods. Studies should

be performed to determine how the distribution changes by store, location, and land type.

Additionally, the data presented here should also be examined utilizing the residuals

method to confirm, with another method, how the data best fit this distribution. Another

analysis, with more data points, should be undertaken to determine the distribution with

better accuracy.

The above experimental data on grocery store foods was determined utilizing

previous studies as a template as well as a guideline to choose which samples to measure.

Brazil nuts, although not in the original work, prove to be hyper accumulators. They

exhibit the highest number of peaks of any measured food. Their shells also exhibit the

same properties. Beef and beef kidney are only one data point and will not be used for the

analysis to follow. An interesting point to note about beef is that it illustrates a low lead-

210 concentration whereas the kidney shows a high lead-210 concentration. The data

examined as a whole illustrate that the concentration of radium-226 is lower, on average,

than lead-210.

This chapter has laid down the concentration and associated distributions of

various radionuclides in specific grocery store foods. The previous chapter determined

dietary intake and the distribution that it is estimated to follow. The next in this series of

chapters deals with the dose conversion factor. The data from these three chapters will be





49


combined in the sixth chapter on dose to provide a more accurate estimate both in

number and shape of distribution than currently available in the literature.













CHAPTER 5
DOSE CONVERSION FACTORS


Introduction

This is the fifth in a series of chapters designed to design and implement a

methodology to determine a probabilistic radiation dose to individuals from foods

bought at local stores in Gainesville, Florida, using the Crystal Ball program. The first

chapter provided the introduction of a probabilistic dose approach. The second chapter

described the literature search for the overall dissertation to determine the references

that were used for each section. The third chapter determined the dietary intake values

to be used and the most probable distribution to describe them. The fourth chapter

described the actual experimentation performed to determine both the concentration of

specific radionuclides in the various foods measured and the distribution to describe this

concentration. The purpose of this chapter is to determine the value and distribution to

describe the dose conversion factor.


Literature Search

A literature search was performed to determine the various applicable references

in an effort to determine the correct dose conversion factor (DCF) as well as a

distribution to apply to it. The data for the dose conversion factor (DCF) came from

four sources. The first source was Federal Regulatory Guide number 11 (EPA, 1988).

This document provides the methodology used to calculate the DCFs for inhalation,








submersion, and ingestion. The tables of the various DCF data for various radionuclides

are included in this manual.

The articles International Council on Radiation Protection (ICRP) 68 (ICRP,

1994) and 72 (ICRP, 1996) provide age-dependent DCFs for workers and members of

the public from intake of radionuclides.

The fourth reference for these data was a solution manual that calculated a dose

conversion factor for strontium 90 (Turner, Bogard, Hunt, & Rhea, 1988, pp. 96-101).

This was utilized as a reference to describe the method to obtain a dose per unit intake

factor from the initial data.

It should be noted that an additional vital source of information describing the

distribution was a direct conversation with Dr. Eckerman at Oak Ridge National

Laboratories. He provided the data that stated that the dose conversion factors follow a

lognormal distribution. Lead-210 and radium-226 distribution encompass 90% of the

values by multiplying and dividing the mean by a factor of five. Polonium-210

distribution can encompass 90% of the values by multiplication and division of the

mean by a factor of 10. These data were incorporated into the final analysis of each

dose analysis as the ninth case.


Discussion

There are two points which need to be addressed at this point: the distribution

and the method to obtain or make an educated estimate and the different sources of

DCFs. Either a distribution can be assumed or it can be calculated utilizing the Crystal

Ball analysis to assign distributions to each of the variables in the equation. Either

approach will produce a final output that will be utilized in the next chapter to calculate








and derive a distribution for the end product dose. The purpose of this chapter is to

derive a probabilistic methodology that can predict a distribution for dose.

There are at least three different sources of dose per unit intake or dose

conversion factors: the EPA Federal Regulatory guide (FRG) number 11, International

Council on Radiation Protection (ICRP) Publication 68 and ICRP 72. These sources and

their DCFs are listed in Table 5-1.


Table 5-1: Dose conversion Factors from Various Sources (Sv/Bq)

ICRP 68 ICRP 72 FRG 11
Pb-210 6.80e-7 6.90e-7 1.450e-6
Ra-226 2.80e-7 2.80e-7 3.58e-7
Po-210 2.40e-7 1.20e-6 5.14e-7


As can be seen from this table, the numbers are not identical and have a rather

large variance. The dose conversion factors from Federal Regulatory Guide Number 11

were utilized in this report. These data were chosen to maintain consistency with

previous reports and to provide comparability with data from those same reports.


Conclusion and Recommendations

The above data for the EPA FRG number 11 will be utilized for the purpose of

this report with the distribution to be assigned as a lognormal distribution. The EPA

FRG 11 dose conversion factors will be converted to mrem/pCi to maintain consistency

and units.

A suggestion for future work is twofold. The discrepancies between the various

agencies and their dose per unit intake should be considered and evaluated.

Additionally, the distribution for this factor should be evaluated utilizing Crystal Ball








and the individual factors in the equation to obtain a more accurate determination of the

distribution for this factor.

The next chapter combines all the previous data and distributions together. The

original 1990 FIPR (Guidry et al., 1990) study is analyzed for a series of distributions

for each of the parameters. Two radionuclides are considered: radium-226 and lead-210.

The resulting Crystal Ball dose distributions are presented in tabular format. A similar

analysis is performed on the grocery store data for each radionuclide. The last chapter

sums all the previous data into a combined whole for comparison and discussion.













CHAPTER 6
COMMITTED EFFECTIVE DOSE EQUIVALENT


Introduction

This chapter designed to determine and test a methodology to calculate a

probabilistic radiation dose to individuals from foods bought at local stores in

Gainesville, Florida. The first chapter provided the introduction of a probabilistic dose

approach. The second chapter described the literature search for the overall dissertation to

determine the references that were used for each section. The third chapter determined

the dietary intake values to be used and the most probable distribution to describe them.

The fourth chapter described the actual experimentation performed to determine both the

concentration of specific radionuclides in the various foods measured and the distribution

to describe this concentration. The fifth chapter described the methodology to determine

the dose conversion factor (DCF) value and distribution. The purpose of this chapter is to

determine the value and distribution to describe the dose to an individual based on the

values obtained in the previous chapters.

This purpose will be accomplished by a literature search that describes the

applicable and relevant literature to determine the committed effective dose equivalent,

the term to describe the extended dose to an individual based on intake of a specific

radionuclide. The data previously obtained were then analyzed with the original data and

comparison is provided.








Literature Search

Dose to an individual can be calculated in several ways. The EPA Federal

Regulatory Guide Number 11 provides the dose conversion factors utilized in this chapter

(EPA, 1988). Other dose conversion factors from ICRP 68 (ICRP, 1994) and ICRP 72

(ICRP, 1996) were considered but not utilized for this analysis. This was to maintain

consistency from the previous 1990 Florida Institute of Phosphate Research (FIPR)

report. These tables allow the user to calculate a dose to an individual based on the

individual's unit intake of a radionuclide.

The 1990 FIPR report was utilized for its diet model and dose analysis of radium-

226 and lead-210 (Guidry et al., 1990). The dose analysis from this report was utilized to

determine the associate dose and distribution to a known and published value for a diet of

an individual living in Florida. The dose model considered from this paper was only the

debris land model for each radionuclide owing to the fact that the maximum individual in

this category had the highest dose.


Method

Determination of dose to an individual is accomplished by integrating the

information determined in the previous papers, multiplying the appropriate factors,

summing, and then running Crystal Ball on the entire set to determine an output.

Equation 1-1 illustrates the formula to calculate dose to an individual. Each variable in

this formula is assigned a value in a Microsoft Excel spreadsheet. The various values for

intake, concentration, and dose conversion factors are assigned a distribution from the

Crystal Ball library. Crystal Ball, produced by Decisioneering, is a program addition to

Microsoft Excel (Decisioneering, 1996). This program utilizes a Monte Carlo sampling








technique for each assigned distribution to determine a final dose in the form of a

distribution.

Monte Carlo is a method in which a random sample is picked from each

distribution and used in the calculation. Each random sampling, with its resultant output,

is called a trial. Depending on the number of trials specified, an output distribution is

framed. The more trials performed, the more accurate the distribution.

Specific fluctuations in the variables such as location of individuals, eating habits,

land type, where food is grown, radiation type and food preparation methods are taken

into account by the distribution determination in each variable. Specific errors are

considered as a whole to contribute to the shape of the distribution. A family of

distributions and analyses are performed to ensure flexibility of the resultant output in

determination of a final dose. Should one specific set not be correct in its distribution

choice, other sets will allow the correct determination of dose and associated

distribution.

Nine sets of analyses were performed on each radionuclide, lead-210, and radium-

226. These analyses were performed on the original 1990 FIPR report and the

experimentally obtained grocery store data. The data are presented in tabular format,

comparing the different distributions uses and the different distributions obtained for each

set. The sets of data are presented below in Table 6-1.

The various sets each have different values for intake, concentration, and dose

conversion factors. LN represents a lognormal distribution, and G represents a gaussian

distribution. The intake values for each variable were obtained from the previous

chapters. The intake values were obtained from the 1990 FIPR report and the third report








in this series. The concentration data were taken from the fourth chapter, the

experimental analyses on the grocery store samples and from the 1990 FIPR report. The

dose conversion factors were obtained from Federal Regulatory Guide # 11. Default

values were utilized for the various parameters that could not be identified. Set nine is a

special case that is similar to set 1 with the exception that the dose conversion factor is

specified by a more exact representation of the actual DCF. Set 9 is the most plausible

scenario for the dose value and distribution. All Crystal Ball analyses were run with

20,000 trials to improve consistency, accuracy, and comparability.


Table 6-1: Sets of Distributions Utilized in Analyses on Data

Intake Concentration DCF
Set 1 LN LN LN
Set 2 G LN LN
Set 3 G G LN
Set 4 G G G
Set 5 LN G G
Set 6 LN LN G
Set 7 LN G LN
Set 8 G LN G
Set 9 LN LN LN


Analysis of 1990 FIPR Dose Diet

Radium 226 Analysis on 1990 FIPR Data

The dose worksheet provided in Table 6-2 shows the spreadsheet for the radium-

226 dose calculation that was used as input to the Crystal Ball program. Each of the

intake variables and concentration variables were assigned a distribution. The dose








Table 6-2: Input Spreadsheet Data for Radium 226 Dose Calculation (Guidry et al., 1990)


DCF


1.30E-03


(mrem/DCi)


Diet Item Intake Concentration Intake
(g/day) (pCi/kg) (pCi/yr)

Broccoli 3.51 34.67 44.42
LEAFY
Cabbage 7.04 32.2 82.74
Collard Greens 0.45 86.23 14.16
Lettuce 23.38 45.41 387.52
Mustard Greens 0.45 64.22 10.55
Spinach 3.28 540.25 646.79
Turnip Greens 0.45 55.47 9.11
SEEDS/GRAINS
Blackeyed Peas 5.61 25.6 52.42
Rice 22.94 82.18 688.10
Yellow Corn 14.41 25.6 134.65
ROOTS______________ _
Carrot 2.92 113.83 121.32
Onion 4.19 33.3 50.93
Radish 0.32 33.3 3.89
Turnip 0.42 23.64 3.62
GENERAL__ __
Cucumber 2.62 18.6 17.79
Green Beans 8.8 9.79 31.45
Green Peppers 1.99 18.6 13.51
Strawberries 1.23 806.68 362.16
Tomato 25.18 18.6 170.95
Watermelon 3.44 18.6 23.35
Squash/Zucchini 1.26 5.15 2.37


rotals


133.89


rotal Diet 1 3071.81


2871.781


Dose
Non-Sampled 2.18E+00 mrenm/yr
Sampled 3.73E+00 mrem/yr
Total 5.92E+00 mrem/yr








conversion factor was also assigned a distribution. The product of these three variables

and a conversion factor allowed the determination of a dose and a distribution.

Table 6-3 shows the output of the family of analyses obtained when the various

distributions were placed into the appropriate variables. Table 6-4 illustrates the statistical

data to allow for a comparison of the various distributions.

Figure 6-1 is the output of the Crystal Ball forecast for set 9. Figure 6-2 shows the

difference comparison between the best fit distribution and the Monte Carlo Crystal Ball

determination of the distribution. The scale on the y-axes illustrates that there was close

agreement.


Table 6-3: Table of Input and Output Distributions for Radium-226


Set 1 LN LN LN LN
Set 2 G LN LN LN
Set 3 G G LN LN
Set 4 G G G G
Set 5 LN G G Beta
Set 6 LN LN G Beta
Set 7 LN G LN LN
Set 8 G LN G Beta
Set 9 LN LConcentration DCF DoseLN




Set 9_ LN LN LN LN


Table 6-4: Statistical Comparison of Various Data for Radium-226


Set 1 5.91 0.623 0.3 3.17 0.11 4.81
Set 2 5.92 0.625 0.32 3.13 0.11 5.16
Set 3 5.92 0.632 0.3 3.15 0.11 5.1
Set 4 5.92 0.629 0.06 3.04 0.11 4.91
Set 5 5.92 0.627 0.05 3.04 0.11 5.12
Set 6 5.92 0.626 0.05 2.97 0.11 4.62
Set 7 5.91 0.624 0.32 3.3 0.11 6.23
Set 8 5.92 0.624 0.08 3.04 0.11 5.28
Set 9 9.56 12.1 5.08 53.59 1.26 270


COF


Intake


DCF


Dose


Mean










Oierlav Mart
Reqjency Comprison

k
.. .. .


MOG -I1

.IO

.(n1oo


112I1 22B1


* kuTdDjrbia.
9 j=d DaI/=1E+1

STB=12l&1

* Tda


33E1 453&1


Figure 6-1: Set 9 Output and Distribution Fit for Ra-226


Figure 6-2: Difference Chart for Ra-226 Set 9


Lead-210 Analysis on 1990 FIPR Data

The lead-210 analyses on the 1990 FIPR data was performed in the same manner as

the radium-226 above. The input data, represented in spreadsheet form, is presented in

Table 6-5 below. Table 6-6 is the presentation or the distribution data for the various


Oera Chart
Freqncy DOffenc


Sun-dDiticn
BDnv=96B-O
adDa&=12l&1l


m Tea


22&E1


QO(-+O


45f&1








Table 6-5: Input Spreadsheet Data for Lead-210 Dose Calculation (Guidry et al., 1990)


DCF


5.40E-03


(mrem/pCi)


Diet Item Intake Concentration Intake
(g/day) (pCi/kg) (pCi/yr)

Broccoli 3.51 60.09 76.98
LEAFY______________ _
Cabbage 7.04 122.61 315.06
Collard Greens 0.45 33.29 5.47
Lettuce 23.38 75.56 644.81
Mustard Greens 0.45 0.50 0.08
Spinach 3.28 166.49 199.32
Turnip Greens 0.45 40.48 6.65
SEEDS/GRAINS
Blackeyed Peas 5.61 22.00 45.05
Rice 22.94 62.26 521.31
Yellow Corn 14.41 22.00 115.71
ROOTS
Carrot 2.92 5.97 6.36
Onion 4.19 4.70 7.19
Radish 0.32 4.70 0.55
Turnip 0.42 10.22 1.57
GENERAL
Cucumber 2.62 8.00 7.65
Green Beans 8.80 8.00 25.70
Green Peppers 1.99 8.00 5.81
Strawberries 1.23 456.19 204.81
Tomato 25.18 8.00 73.53
Watermelon 3.44 8.00 10.04
Squash/Zucchini 1.26 8.00 3.68


Totals


133.89


Total Diet 1 3071.81


2277.31957


Dose
Non-Sampled 9.07E+00 mrem/yr
Sampled 1.23E+01 mrem/yr
Total 2.14E+01 mrem/yr








Table 6-6: Table of Input and Output Distributions for Lead-210


Dose


DCF


Intake


Concentration


Set 1 LN LN LN LN
Set 2 G LN LN LN
Set 3 G G LN LN
Set 4 G G G Beta
Set 5 LN G G Beta
Set 6 LN LN G G
Set 7 LN G LN LN
Set 8 G LN G Beta
Set 9 LN LN LN LN


variables and the final dose. Table 6-7 shows the statistics of interest for the output

distributions.

Figure 6-3 is shown as the set 9 for this family of distribution analyses. It is

believed to be the most probable outcome of dose. Figure 6-4 shows the Crystal Ball

output graph of the difference chart. This graph illustrates the difference between the

output dose distribution and the closest fit approximation determined by the program and

its subsequent Chi-squared test of fit for the curve.


Table 6-7: Statistical Comparison of Various Data for Lead-210


Mean


Std Dev


Skewness


Kurtosis


COF


Range
Width


Set 1 21.4 2.25 0.32 3.18 0.11 17.6
Set 2 21.4 2.25 0.29 3.1 0.11 18.2
Set 3 21.3 2.25 0.37 3.27 0.11 19.6
Set 4 21.3 2.25 0.06 3.04 0.11 18.4
Set 5 21.4 2.27 0.05 3.06 0.11 19
Set 6 21.3 2.28 0.08 2.99 0.11 19
Set 7 21.4 2.25 0.33 3.18 0.11 18.2
Set 8 21.4 2.28 0.07 3.05 0.11 18.1
Set 9 34.7 43.9 5.18 56.66 1.25 965









Oeray Choart
Freqancymnpaison


.0.
I'

.011 I

.co a
.CM O


3-1I 76&1


W g=darb cn
Man=34&1
SUDBe=43E1


*TCW


1.122 1&2


Figure 6-3: Set 9 Output and Distribution Fit for Pb-210


OieaChart
Frecpency Uffermice


.0CI J --dDri,=431 '1

-I B --Ta


OBO 31 7aE1 1.1-2 1BE-2

Figure 6-4: Difference Chart for Lead-210 Set 9


Grocery Store Data Analysis

Lead-210

The analyses on the grocery store data followed the same approach and

methodology as that utilized to perform the analyses on the original FIPR data. Two

radionuclides were considered: lead-210 and radium-226. In fact, like the original study,








this study counted the foods for two radionuclides: lead-210 and radium-226. The

original study provided the data but no analysis on the polonium due to the lack of

literature and values for the nonsampled food items and the concentration or dose that

could be assigned to these values. A literature search did not turn up enough data to

support analysis of this third radionuclide; therefore, similar to the original study, no

analyses were performed in relation to it.

Each variable is assigned a value. Intake data were obtained from a literature

search of previous studies and databases on consumption in the United States. The

concentration data for the various foods was obtained by experimental measurement

discussed in the fourth chapter. Dose conversion factors were obtained from Federal

Regulatory Guide # 11. All of these data were provided different distributions to obtain

the nine sets of dose distributions and analyses. Table 6-8 shows the input to Crystal

Ball.

Table 6-9 illustrates the input variable and the output distributions obtained for

the various sets tested. Table 6-10 shows the statistical output for the various sets. Figure

6-5 is the forecast output of the Crystal Ball program for the 9th set, and Figure 6-6 shows

the difference between the programs best fit and the output data.


Radium-226

Table 6-11 illustrates the data input to the program to run the simulation and

obtain results. Table 6-12 is the description of the distribution results. Table 6-13 shows

the output statistics. Figure 6-7 is the frequency output for set 9, and Figure 6-8 is the

difference comparison of this frequency output to the best fit distributions.









Table 6-8: Input Spreadsheet Data for Lead-210 Dose Calculation


I DCF


I 5.40E-03 I (mrem/pCi)


Diet Item Intake Concentration Intake
(g/day) (pCi/kg) (pCi/yr)
Broccoli 3.51 829 1062.07
LEAFY___________
Cabbage 7.04 858 2204.72
Collard Greens 0.45 1372 225.35
Lettuce 23.38 1237 10556.19
Mustard Greens 0.45 1622 266.41
Spinach 3.28 422 505.22
Turnip Greens 0.45 941 154.56
SEEDS/GRAINS
Blackeyed Peas 5.61 494 1011.54
Rice 22.94 654 5476.01
Yellow Corn 14.41 506 2661.38
ROOTS
Carrot 2.92 555 591.52
Onion 4.19 599 916.08
Radish 0.32 580 67.74
Turnip 0.42 401 61.47
GENERAL _____
Cucumber 2.62 47082 45024.52
Green Beans 8.80 3837 12324.44
Green Peppers 1.99 431 313.06
Strawberries 1.23 715 321.00
Tomato 25.18 1011 9291.80
Watermelon 3.44 776 974.35
Squash/Zucchini 1.26 613 281.92


Totals


133.89


Total Diet 3071.81

Dose
Non-Sampled 9.07E+00 mrem/yr
Sampled 5.09E+02 mrenim/yr
Total 5.18E+02 mrem/yr


94291.34








Table 6-9: Table of Input and Output Distributions for Lead-210


1 1


Set 1 LN LN LN LN
Set 2 G LN LN Gamma
Set 3 G G LN Gamma
Set 4 G G G Gamma
Set 5 LN G G Gamma
Set 6 LN LN G Gamma
Set 7 LN G LN LN
Set 8 G LN G Gamma
Set 9 LN LN LN LN


Table 6-10:


Statistical Comparison of Various Dose Rate Results Data for Lead-210


Mean


Std Dev


Skewness


Kurtosis


COF


Range
Width


Set 1 518 63.9 0.41 3.26 0.12 563
Set 2 518 64.6 0.36 3.19 0.12 529
Set 3 518 64.2 0.37 3.23 0.12 541
Set 4 518 64.7 0.2 3.08 0.12 588
Set 5 518 63.9 0.21 3.12 0.12 541
Set 6 517 63.8 0.22 3.14 0.12 576
Set 7 518 64.6 0.39 3.26 0.12 529
Set 8 518 63.7 0.21 3.16 0.12 529
Set 9 854 1070 6.02 86.32 1.27 3150


Figure 6-5: Set 9 Output and Distribution Fit for Pb-210


Concentration


Intake


Freeny Compason


.lM 1.-3




.ODO
"I.O 0B3 201d3 acEE*3 400B3


DCF


Dose


























Figure 6-6: Difference Chart for Lead-210 Set 9


Overall Analysis

The above sets were each run with 20,000 trials each to decrease statistical error

and improve comparability. The data show some very interesting results when compared

to each other and then when compared to overall annual dose to an individual.


Dose Data Comparison

Lead-210 exhibited the highest dose in both the original set of data and the

grocery store data. The original FIPR 1990 report data yielded a dose from

ingestion of lead 210 of 21.4 mrem/yr for sets 1 through 8 and 34.7 mrem/yr

for set 9. The grocery store data yielded a dose to the individual of 518 mrem/yr for

sets 1 through 8 and 854 mrem/yr for set 9 individual of 877mrem/yr. The dose

calculated from the grocery store data was 24 times higher than that calculated from the

1990 FIPR data.


Qlerly Chart
Freency fferenem

I'!-l--1-"--"r''"-
I- mTElm
I H ||, | MEn=a45E&2
l I I I Lu ll I I -.SC/10&
U0 T TlD rW
-""* -1 --- 1-------------**2


OCEO taE&3 2IBE+3 3IE+3 4aB3








Table 6-11: Input Spreadsheet Data for Radium-226 Dose Calculation (Grocery Data)


-I. r


Diet Item Intake Concentration Intake
______(g/day) (pCi/kg) (pCi/yr)

Broccoli 3.51 5.00 6.406
LEAFY______ _
Cabbage 7.04 6.00 15.418
Collard Greens 0.45 6.00 0.986
Lettuce 23.38 4.00 34.135
Mustard Greens 0.45 17.00 2.792
Spinach 3.28 20.00 23.944
Turnip Greens 0.45 6.00 0.986
SEEDS/GRAINS
Blackeyed Peas 5.61 3.00 6.143
Rice 22.94 2.00 16.746
Yellow Corn 14.41 6.00 31.558
ROOTS__________________ _
Carrot 2.92 4.00 4.263
Onion 4.19 4.00 6.117
Radish 0.32 4.00 0.467
Turnip 0.42 1177.00 180.434
GENERAL___ ___
Cucumber 2.62 9.00 8.607
Green Beans 8.80 7.00 22.484
Green Peppers 1.99 4.00 2.905
Strawberries 1.23 4.00 1.796
Tomato 25.18 3.00 27.572
Watermelon 3.44 2.00 2.511
Squash/Zucchini 1.26 4.00 1.840


totals


133.89


rotal Diet I 3071.81


398.109


Dose
Non-Sampled 2.18E+00 mrem/yr
Sampled 5.18E-01 mrem/yr
Total 2.70E+00 m/rem/yr


(mrem/vCi)


1.30E-03


DCF








Table 6-12: Table of Input and Output Distributions for Radium-226 (Grocery Data)


Intake


Dose


Concentration


DCF


Set 1 LN LN LN LN
Set 2 G LN LN Gamma
Set 3 G G LN Gamma
Set 4 G G G Beta
Set 5 LN G G Beta
Set 6 LN LN G Beta
Set 7 LN G LN Gamma
Set 8 G LN G Beta
Set 9 LN LN LN LN


Table 6-13: Statistical Comparison of Various Dose Rate Results Data for Radium-226
(Grocery Data) -___


Mean


Std Dev


Skewness


Kurtosis


COF


Range
Width


Set 1 2.7 0.273 0.31 3.22 0.1 2.27
Set 2 2.7 0.275 0.31 3.29 0.1 2.25
Set 3 2.7 0.274 0.31 3.15 0.1 2.06
Set 4 2.7 0.27 0.02 2.95 0.1 2.1
Set 5 2.7 0.273 0 2.95 0.1 2.21
Set 6 2.7 0.272 0.04 3.03 0.1 2.16
Set 7 2.7 0.271 0.3 3.21 0.1 2.45
Set 8 2.7 0.272 -0.01 3.02 0.1 2.09
Set 9 4.38 5.52 5.6 69.91 1.28 150


ktw=43Eto
*





QOE1O 5(IE0O tE+1 1.5E+1 2CIE-1



Figure 6-7: Set 9 Output and Distribution Fit for Radium-226 (Grocery Data)


Freqemrncy Cmaison






















Figure 6-8: Difference Chart for Radium-226 Set 9 (Grocery Data)


The dose from the grocery store data is believed to be higher for two reasons. The

peak measured on the original analysis was the 40 KeV peak. The peak measured on the

grocery store analysis was the 10.8 KeV peak. This peak was measured because of the

successful recognition of this peak by the GammaVision program for this peak for the

lead-210 radionuclide. This 10.8 KeV peak yielded a significantly higher correction

factor than previously measured and observed for the alternate lead-210 peak.

Radium-226 had the lowest dose per year of both radionuclides considered. The

output from the analysis of the 1990 FIPR data yielded a dose to the individual from

radium-226 of 5.91 mrem/yr for sets 1 through 8 and 9.56 for set 9. The grocery store

data provided an individual dose that was lower than the FIPR data. The grocery store

data dose to the individual was calculated to be 2.7 mrem/yr for sets 1-8 and 4.38 for set

9. The dose calculated from the grocery store data was 45% of the dose calculated for the

original FIPR dose calculation.


Frequency Off Wears


I'I il' r Ile = 3&
U ,I I~l, ,,' i. - SC =5B
-U I T D't
rM ---
.CMl La n_


001-O 5 -BO O1.0 1 1.-1 2(]]1








Comparing the dose to the individual from radium-226 and lead-210 from the

FIPR data, it was calculated that dose from lead-210 was 3.6 times higher than the

individuals dose from radium-226. A similar comparison of doses for the grocery store

data concluded that the individual's dose attributable to lead-210 was 191 times greater

than that attributable to radium-226

Some of the differences between doses attributable to the radionuclides can be

traced back to their reported dose conversion factors. Lead-210 is almost three times

larger than radium-226. Consider this factor with the fact that lead-210 had some

individual samples with high counts and a higher intersample comparison should be

expected.


Comparison of Distributions and Statistics

The dose conversion factor distribution appears to have a strong effect on the dose

distribution. A sensitivity analysis was performed on the various variables with the

outcome determining that the dose conversion factor was over 10 times more sensitive to

the output distribution that any other variable. The data reveal that multiplying three

lognormal distribution yields a lognormal output distribution each time. Multiplying three

normal distributions times each other provides a normal distribution for the output dose

in two cases and a beta distribution in two other cases. Multiplying a mixture of

lognormal distributions and normal distributions does not necessarily yield either of

these distributions as an output. The best fit to the dose distribution output in the

preceding case is illustrated above to be a lognormal, a normal, a beta, or a gamma

distribution.








The beta distribution is a very flexible distribution commonly utilized to represent

variability over a wide range (Decisioneering, 1996). This distribution can assume a wide

variety of shapes when the values of alpha a beta are varied.

The gamma distribution is related to the lognormal distribution and is used

sometimes to represent pollutant concentrations and precipitation quantities.

Set 9, in all four sets of analyses, was utilized to represent the most accurate

quantity. The value of this dose in all four groups was significantly higher that the other

eight sets in that group under consideration. The primary reason for this is the change in

the dose conversion factor. Being the most sensitive variable and the large skewness

introduced due to the determination of an accurate shape of this curve from a discussion

with Dr. Eckerman, this provided an output dose distribution different from the other sets.


Conclusions

The purpose of this chapter was accomplished by a literature search that described

the applicable and relevant literature to determine the committed effective dose

equivalent, the term to describe the extended dose to an individual based on intake of a

specific radionuclide. The data previously obtained were then analyzed with the original

data and a comparison is provided.

The shape of various dose distributions to individuals was determined for nine

combinations of variable distributions in each of the two radionuclides in each of the two

studies. The data from one study came from a previous 1990 FIPR report. The data for

the other study were determined by experimental measurement from an earlier chapter in

this series. The Crystal Ball program and monte carlo sampling method inherent to it

were utilized to perform these analyses for each set of distributions.








The average individual annual dose, as stated in the introduction, is 360 mrem/yr.

Lead-210 measurement is 518 mrem/yr for eight sets of analyses on grocery store

samples and 854 mrem/yr for the most likely distribution scenario. The lead-210 dose to

an individual from the analysis of the FIPR data was 34.7 mrem/yr for the case 9 and

21.4 mrem/yr for all others. The case 9 set of variable distributions is believed to be the

most likely dose output. This output distribution is described by a lognormal distribution.

Many reasons were observed for the higher value from the experimental data dose

determination. A different peak measurement and a higher correction factor were both

significant factors as well as the fact that the 10.8 KeV is at the lower edge of analysis for

the system. In defense of a higher lead-210 dose, the original 1990 FIPR report chose not

to use grocery store data due to the fact that only single replicates were being measured

and the alternate fact that the dose obtained was 200 times greater than the other

measurements.

The radium-226 individual dose measurements for the case 9, the most likely

case, also followed a lognormal distribution. The values of the dose to the individual

were significantly lower than the lead-210 values. The dose for the FIPR data was 5.91

mrem/yr for eight cases and 9.56 mrem/yr for the ninth and most likely case. The grocery

store data provided a dose to the individual that measured 2.7 mrem/yr for 8 sets and 4.38

mrem/yr for the most likely case. These data, compared to the 360 mrem/yr expected

dose from natural sources, are less than 3% for all radium-226 measurements.

The distributions described by the various analyses included normal, lognormal,

beta, and gamma. The output, dose, and distributions were strongly influenced by the

distribution assigned to the dose conversion factor variable.








This chapter is designed to determine a probabilistic dose to individuals from

foods bought at local stores in Gainesville, Florida. The first chapter provided the

introduction of a probabilistic dose approach. The second chapter described the literature

search for the overall dissertation to determine the references that were used for each

section. The third chapter determined the dietary intake values to be used and the most

probable distribution to describe them. The fourth chapter described the actual

experimentation performed to determine both the concentration of specific radionuclides

in the various foods measured and the distribution to describe this concentration. The

fifth chapter described the methodology to determine the dose conversion factor (DCF)

value and distribution. The purpose of this chapter was to determine the value and

distribution to describe the dose to an individual based on the values obtained in the

previous chapters. The next and final chapter provides a conclusion and

recommendations for future research in this area.













CHAPTER 7
CONCLUSIONS AND RECOMMENDATIONS


This is the seventh and the last chapter intended to design and implement a

methodology to determine a probabilistic radiation dose to individuals. This methodology

is analyzed by application to data experimentally determined from foods bought at local

stores in Gainesville, Florida.

The first chapter in the series provided the introduction of a probabilistic dose

approach. The explanation of how the current approach to individual dose determination

was deterministic and resulted in a dose with a single number describing it. A

probabilistic approach was explained and how this yielded a dose that was described with

a distribution.

The second chapter described the literature search for the overall dissertation to

determine the references that were used for each section. This general literature search

provided the basis for the background data utilized for the remainder of the dissertation.

The third chapter determined the dietary intake values to be used and the most

probable distribution to describe them. These values and data were derived from literary

sources of previous studies and dietary surveys performed in the United States.

The fourth chapter described the actual experimentation performed to determine

both the concentration of specific radionuclides in the various foods measured and the

distribution to describe this concentration. Samples from three different local stores were

analyzed to determine concentrations of radium-226 and lead-210 in the foods. An








additional study was performed to determine experimentally the type of distribution to

describe best the concentration in foods.

The fifth chapter described the methodology to determine the dose conversion

factor (DCF) value and distribution. Literary sources as well as direct conversation with

people at the cutting edge of dose conversion factor determination provided the data

obtained in this chapter for the value and distributions to describe the dose conversion

factor.

The sixth chapter determined the value and distribution to describe the dose to an

individual based on the values obtained in the previous chapters. The Crystal Ball

program was utilized to perform groups of analyses on both original 1990 Florida

Institute of Phosphate Research data as well as grocery store data.

Crystal Ball is like any computer program. Good information in will yield good

information out. The data that were measured had numerous data points that provided no

counts at the specific peaks that were being observed. Minimum detectable activity

numbers were input for these data points to provide for a large and, therefore

conservative from a safety standpoint, dose estimate. The unsampled data points provided

another source of error which was overcome by reference to previous work and data

manipulation to provide a dose for these data points. This, however, provides another

possible source for higher dose output. The Crystal Ball program provided an accurate

way to sample the input to provide a distribution output that was as accurate as the input

provided.

The grocery store data showed a lognormal distribution for lead-210 with an

average dose 854 mrem/yr for the most probable case and 518 mrem/yr for all other eight








cases. The FIPR data illustrated a lead-210 mean of 34.7 mrem/yr for the most probable

case and 21.4 consistently for all other cases.

The reasons for such a high lead-210 dose is due to many factors: different peak

counted, different calibration or correction factor, and the peak measured being at the

lower edge of the detector's measurement limit. This number, as stated before, seems

abnormally high and should be verified with other analyses.

The grocery store radium-226 measurement for individual dose, by comparison,

was 2.7 mrem/yr for eight cases and 4.38 mrem/yr for the ninth and most probable case.

This value was lower than the individual dose value obtained from the data from the 1990

FIPR report. This report yielded an individual dose of 5.91 mrem/yr for eight cases and

9.56 mrem/yr for the most probable case. The most probable dose scenarios followed a

lognormal distribution that closely approximated the dose conversion factor input

distribution.

The numbers for radium-226 are very similar to those obtained by FIPR in the

1990 report and so support their data as well as extending the database. The lead-210

measured in this report are consistent with the previous report in the fact that the lead-210

general foods category radiation dose calculation from the grocery store samples ranged

from 2 to 200 times higher than the literature value (Guidry et al., 1990).

The methodology of using Crystal Ball and probabilistic dose calculation was

successfully implemented in the previous chapter. The inclusion of distributions to

account for fluctuations, errors and variations was very useful and provided a higher

initial dose than the deterministic approach in the previous FIPR paper (Guidry et al.,

1990). The distribution groups allow for flexibility should future research determine more








accurate distributions for the variables. Additionally, the output distributions provide a

visual aid for the public and researchers to understand radiation dose and the fact that it is

a range and not just one number!

Future analyses should, if possible, include the location of where samples are

grown. Additional measurements should be taken to determine and confirm the normal

distribution assigned to the food concentration. Additional Crystal Ball analysis should be

performed on the calculational parameters and, therefore, the solution for the dose

conversion factor formula. This would provide a more accurate distribution for the dose

per unit intake factor. An analysis to determine the differences and advantages of the

various dose conversion factors should be considered. A study should also be undertaken

to determine the effect on changing the various distributions on the final dose. A

determination should be made as to what may be the cause for the elevated measurements

for lead-210. These are a few suggestions on the direction that this work should take in

the future.














APPENDIX A
DATA SHEETS












Item Store Weight Time Counted Detector File Saved
Beef Publix 556.6 34187.1 4 beef.spc
Beef Kidney Publix 449.2 86362.4 4 bfkid.spc
Black-Eyed Peas Publix 455.35 86437.5 2 bleypeas
Black-Eyed Peas Albertsons 264.3 72000 2 bepe.spc
Black-Eyed Peas Winn Dixie 473.8 34098.9 4 bley3.spc
Brazil Shells Publix 304.2 43149 2 bznutssh.spc
Broccoli Publix 267 34167.9 4 broccl.spc
Broccoli Albertsons 192.4 71946.1 2 broc2.spc
Broccoli Winn Dixie 324.4 34124.9 2 broc3.spc
Cabbage Publix 210.45 34187.4 2 cabb.spc
Cabbage Albertsons 179.3 34155.1 4 cabb2.spc
Cabbage Winn Dixie 214.15 34122.6 4 cabb3.spc
Carrots Publix 286.9 34187.8 2 carr.spc
Carrots Albertsons 421.1 71940.8 4 carr2.spc
Carrots Winn Dixie 365.9 34061.3 4 carr3.spc
Cauliflower Publix 268.8 34187.4 4 caul.spc
Cauliflower Albertsons 154.45 34187.5 2 caul2.spc
Cauliflower Winn Dixie 396.1 34167.7 2 caul3.spc
Collard Greens Publix 140.35 34187.9 4 cogr.spc
Collard Greens Albertsons 421.15 71940.8 2 cogr2.spc
Collard Greens Winn Dixie 136.6 34104.7 2 cogr3.spc
Corn Publix 289.85 34194.1 2 corn.spc
Corn Albertsons 199.6 34187.3 4 corn2.spc
Corn Winn Dixie 418.4 34051.1 4 corn3.spc
Cucumber Publix 347.55 34167.9 2 cucul.spc
Cucumber Albertsons 257.6 71942.7 2 cucu2.spc
Cucumber Winn Dixie 324.2 34087.2 2 cucu3.spc
Eggplant Publix 238.4 34187.8 2 eggp.spc
Eggplant Albertsons 231.7 71970.2 4 egpl2.spc
Eggplant Winn Dixie 130.3 34089 2 eggp3.spc
Grapefruit Publix 448.1 34186.8 4 grap.spc
Grapefruit Albertsons 317.6 71944 4 grap2.spc
Grapefruit Winn Dixie 422.5 34089 4 grap3.spc









Item Store Weight Time Counted Detector File Saved
Green Beans Publix 216.7 34186.9 4 grbr.spc
Green Beans Albertsons 250 71971.9 2 grbe2.spc
Green Beans Winn Dixie 338 34036.9 2 grbe3.spc
Green Onions Publix 97.2 34187.8 4 gron.spc
Green Onions Albertsons 89.1 71972.1 4 gron2.spc
Green Onions Winn Dixie 195.15 34100.7 2 gron3.spc
Green Peppers Publix 321 34186.9 2 grpe.spc
Green Peppers Albertsons 287.5 71941.8 4 grpe2.spc
Green Peppers Winn Dixie 415.75 34097.7 4 grpe3.spc
Irish Creamer Potatoes Publix 353.75 34186 4 icp.spc
Lemons Publix 316.9 43176.9 4 lem.spc
Lemons Albertsons 238.75 71941.8 2 lemo2.spc
Lemons Winn Dixie 372.7 34154.8 4 lemo3.spc
Lettuce Publix 218.8 43181.5 2 lettw.spc
Lettuce Albertsons 201.1 71944 2 lett2.spc
Lettuce Winn Dixie 319.45 34099.5 2 lett3.spc
Lima Beans Publix 362.7 43182.2 2 lima.spc
Lima Beans Albertsons 358.65 71942.7 4 libe2.spc
Lima Beans Winn Dixie 382.6 34040.6 2 libe3.spc
Mustard Greens Albertsons 81.35 34187.5 4 mugr.spc
Okra Publix 180.5 34186 2 okra.spc
Okra Albertsons 240.5 71970.2 2 okra2.spc
Okra Winn Dixie 320.3 34167.4 4 okra4.spc
Onions Publix 335.6 43181.9 2 onio.spc
Onions Albertsons 323.15 71979.8 2 onio2.spc
Onions Winn Dixie 332.1 34037.1 4 onio3.spc
Oranges Publix 446.9 43176.9 2 oran.spc
Oranges Albertsons 381.3 71973.3 2 oran2.spc
Oranges Winn Dixie 379.6 34030.4 4 oran3.spc
Parsley Publix 71.25 43179.7 4 pars.spc
Parsley Albertsons 17.8 34181 2 pars2.spc
Parsley Winn Dixie 175.75 34062 2 pars3.spc




U


Item Store Weight Time Counted Detector File Saved
Peas Publix 263.2 43180.5 2 peas.spc
Peas Albertsons 289.35 71970.9 4 peas2.spc
Peas Winn Dixie 342.6 33922.1 2 peas3.spc
Pole Beans Publix 267.15 34186.8 2 pb.spc
Pole Beans Winn Dixie 416.5 34057.6 4 pobe3.spc
Potato Publix 398 43179.7 2 pot.spc
Potato Albertsons 254.45 86366.8 2 pot2.spc
Potato Winn Dixie 390.3 34154.9 2 pot3.spc
Purple Hull Peas Publix 391.05 43182.3 4 php.spc
Purple Hull Peas Winn Dixie 430.3 34200 2 crpe3.spc
Radishes Publix 333.7 34128.8 4 rad.spc
Radishes Albertsons 170.1 71968.7 4 radi2.spc
Radishes Winn Dixie 310.3 34100.9 4 rad3.spc
Red Potatoes Publix 348.15 43182.2 4 rpot.spc
Red Potatoes Winn Dixie 371.8 34086.4 4 rpot3.spc
Rice Publix 518.55 43182 4 rc.spc
Rice Albertsons 446.8 71968.6 2 rice2.spc
Rice Winn Dixie 583.85 34078.7 2 rice3.spc
Spinach Publix 99.85 34185.9 4 sp.spc
Spinach Albertsons 69.3 22702.8 4 spin2.spc
Spinach Winn Dixie 484.7 34077.8 4 spin3.spc
Strawberries Publix 348.75 34126.7 2 stra.spc
Strawberries Albertsons 278.5 71973.3 4 straw2.spc
Strawberries Winn Dixie 337.1 34051.8 2 straw3.spc
Swiss Chard Publix 173.65 34187.1 2 swch.spc
Tangerine Albertsons 276 34155.2 2 tan2.spc
Tangerine Winn Dixie 399.85 34109.5 4 tan3.spc
Tomatoes Publix 358.8 43181.4 4 tomat.spc
Tomatoes Albertsons 376.05 34191.8 2 tom2w.spc
Tomatoes Winn Dixie 434.4 34048 2 5om3.spc
Turnip Greens Publix 135.65 43183.1 4 tg.spc
Turnip Greens Albertsons 313.65 34181.8 4 tugr2.spc
Turnip Greens Winn Dixie 564.2 34185.4 2 tugr3.spc









Item Store Weight Time Counted Detector File Saved
Turnip Root and Green Publix 189.65 34186 2 trag.spc
Turnip Roots Publix 275.1 43148.7 4 turnroots.spc
Turnip Roots Albertsons 0.3 71972.2 2 turn2.spc
Watermelon Publix 463.6 43182.4 2 wm.spc
Watermelon Albertsons 396 34187.4 2 wat2.spc
Yellow Corn Publix 324.6 43182.1 2 yc.spc
Yellow Corn Albertsons 234.75 34191.7 4 yc2.spc
Yellow Corn Winn Dixie 386.6 34167.8 2 yeco3.spc
Yellow Squash Publix 331.6 43183.1 2 yesq.spc
Yellow Squash Albertsons 219.4 86366.8 4 yesq2.spc
Yellow Squash Winn Dixie 361.9 33914.5 4 yellow3.spc
Zucchini Publix 394.3 43180.5 4 zucc.spc
Zucchini Albertsons 256.4 71971.9 4 zucc2.spc
Zucchini Winn Dixie 368.7 34167.8 4 zucc3.spc
Brazil Nuts ______226 86369 4 bznuts.spc










Pb-210 Peaks


Pb-210


10.8 KeV


Item Net (Counts) Error
Beef 0 0
Beef Kidney 126 47
Black-Eyed Peas 0 0
Black-Eyed Peas 0 0
Black-Eyed Peas_____
Brazil Shells 0 0
Broccoli 0 0
Broccoli 0 0
Broccoli 0 0
Cabbage 0 0
Cabbage 0 0
Cabbage 45 12
Carrots 0 0
Carrots 91 45
Carrots 0 0
Cauliflower 105 27
Cauliflower 0 0
Cauliflower 0 0
Collard Greens 43 26
Collard Greens 0 0
Collard Greens 0 0
Cornm 0 0
Corn 0 0
Corn 0 0
Cucumber 4779 142
Cucumber 182 35
Cucumber 0 0
Eggplant 0 0
Eggplant 0 0
Eggplant 0 0
Grapefruit 0 0
Grapefruit 0 0
Grapefruit 0 0









I Pb-210 Peaks


Pb-210


10.8 KeV


Item Net (Counts) Error
Green Beans 437 59
Green Beans 211 37
Green Beans 0 0
Green Onions 0 0
Green Onions 0 0
Green Onions 0 0
Green Peppers 0 0
Green Peppers 0 0
Green Peppers 0 0
Irish Creamer Potatoes 0 0
Lemons 0 0
Lemons 0 0
Lemons 0 0
Lettuce 0 0
Lettuce 0 0
Lettuce 0 0
Lima Beans 76 0
Lima Beans 76 47
Lima Beans 0 0
Mustard Greens 41 29
Okra 0 0
Okra 0 0
Okra 28 8
Onions 0 0
Onions 0 0
Onions 0 0
Oranges 0 0
Oranges 100 38
Oranges 0 0
Parsley 0 0
Parsley 0 0
Parsley 0 0









Pb-210 Peaks
Pb-210 10.8 KeV
Item Net (Counts) Error
Peas 0 0
Peas 49 44
Peas 0 0
Pole Beans 0 0
Pole Beans 0 0
Potato 0 0
Potato 0 0
Potato 0 0
Purple Hull Peas 0 0
Purple Hull Peas 0 0
Radishes 19 30
Radishes 143 43
Radishes 26 7
Red Potatoes 0 0
Red Potatoes 37 12
Rice 0 0
Rice 122 39
Rice 0 0
Spinach 0 0
Spinach 0 0
Spinach 0 0
Strawberries 0 0
Strawberries 0 0
Strawberries 0 0
Swiss Chard 0 0
Tangerine 0 0
Tangerine 0 0
Tomatoes 0 0
Tomatoes 0 0
Tomatoes 87 18
Turnip Greens 69 33
Turnip Greens 0 0
Turnip Greens 79 20









Pb-210 Peaks


Pb-210


10.8 KeV


Item Net (Counts) Error
Turnip Root and Green 0 0
Turnip Roots 53 33
Turnip Roots 0 0
Watermelon 0 0
Watermelon 0 0
Yellow Corn 0 0
Yellow Corn 0 0
Yellow Corn 0 0
Yellow Squash 0 0
Yellow Squash 135 44
Yellow Squash 0 0
Zucchini 0 0
Zucchini 0 0
Zucchini 0 0
Brazil Nuts 235 86










~~____~_____Ra-226 Peaks____________


Pb-214 I295KeV1 Pb-214


352 KeV


Bi-214


609 KeV


Item Net (Counts) Error Net (Counts) Error Net (Counts) Error
Beef 0 0 0 0 0 0
Beef Kidney 0 0 0 0 271 30
Black-Eyed Peas 0 0 0 0 0 0
Black-Eyed Peas 0 0 0 0 0 0
Black-Eyed Peas_________________________
Brazil Shells 495 42 904 43 541 31
Broccoli 0 0 0 0 0 0
Broccoli 0 0 0 0 0 0
Broccoli 39 32 0 0 69 29
Cabbage 0 0 0 0 0 0
Cabbage 96 32 248 30 233 22
Cabbage 0 0 82 29 0 0
Carrots 0 0 0 0 0 0
Carrots 140 44 181 42 0 0
Carrots 0 0 0 0 95 20
Cauliflower 0 0 0 0 0 0
Cauliflower 0 0 0 0 0 0
Cauliflower 0 0 0 0 0 0
Collard Greens 0 0 81 25 0 0
Collard Greens 0 0 0 0 0 0
Collard Greens 0 0 0 0 0 0
Corn 0 0 0 0 0 0
Corn 0 0 0 0 0 0
Corn 0 0 0 0 0 0
Cucumber 0 0 0 0 0 0
Cucumber 51 31 0 0 0 0
Cucumber 0 0 0 0 0 0
Eggplant 0 0 0 0 0 0
Eggplant 75 44 271 37 237 28
Eggplant 0 0 0 0 0 0
Grapefruit 0 0 0 0 0 0
Grapefruit 0 150 39 0 0 0
Grapefruit 0 0 102 26 80 23









Ra-226 Peaks__ _____


Ph.214 [295 KeVI Pb-214 I 352 KeV Bi-214 609 KeV


Item Net (Counts) Error Net (Counts) Error Net (Counts) Error
Green Beans 0 0 0 0 0 0
Green Beans 0 0 0 0 0 0
Green Beans 0 0 0 0 0 0
Green Onions 0 0 0 0 0 0
Green Onions 0 0 227 38 0 0
Green Onions 0 0 0 0 0 0
Green Peppers 0 0 0 0 0 0
Green Peppers 0 0 387 411 303 30
Green Peppers 0 0 0 0 0 0
Irish Creamer
Potatoes 0 0 37 0 0 0
Lemons 0 0 0 67 27 0
Lemons 0 0 0 0 0 0
Lemons 0 0 0 0 6 7
Lettuce 0 0 0 0 0 0
Lettuce 0 0 0 0 0 0
Lettuce 0 0 0 0 0 0
Lima Beans 0 0 299 0 0 0
Lima Beans 180 43 299 40 0 0
Lima Beans 0 0 0 0 0 0
Mustard Greens 0 0 189 26 0 0
Okra 0 0 0 0 0 0
Okra 0 0 0 0 0 0
Okra 0 0 0 0 0 0
Onions 0 0 0 0 0 0
Onions 0 0 0 0 0 0
Onions 0 0 0 0 0 0
Oranges 0 0 0 0 0 0
Oranges 0 0 0 0 0 0
Oranges 0 0 0 0 0 0
Parsley 0 0 100 29 0 0
Parsley 0 0 0 0 0 0
Parsley 0 0 0 0 0 0









II Ra-226 Peaks


I Ph~21S I295K~Vl Ph-214


352 KeV Bi-214


609 KeV


Item Net (Counts) Error Net (Counts) Error Net (Counts) Error
Peas 0 0 0 0 0 0
Peas 0 0 163 38 0 0
Peas 0 0 0 0 0 0
Pole Beans 0 0 0 0 0 0
Pole Beans 0 0 75 24 0 0
Potato 0 0 0 0 0 0
Potato 0 0 0 0 0 0
Potato 0 0 0 0 0 0
Purple Hull Peas 66 33 0 0 0 0
Purple Hull Peas 0 0 0 0 0 0
Radishes 0 0 0 0 0 0
Radishes 0 0 203 37 187 27
Radishes 0 0 0 0 0 0
Red Potatoes 0 0 0 0 0 0
Red Potatoes 0 0 0 0 0 0
Rice 0 0 0 0 0 0
Rice 0 0 0 0 0 0
Rice 0 0 0 0 0 0
Spinach 0 0 0 0 0 0
Spinach 0 0 0 0 0 0
Spinach 0 0 0 0 0 0
Strawberries 0 0 0 0 0 0
Strawberries 0 0 0 0 242 25
Strawberries 0 0 0 0 0 0
Swiss Chard 0 0 0 0 0 0
Tangerine 0 0 0 0 0 0
Tangerine 0 0 0 0 114 15
Tomatoes 0 0 0 0 100 20
Tomatoes 0 0 0 0 0 0
Tomatoes 0 0 0 0 0 0
Turnip Greens 0 0 156 29 104 18
Turnip Greens 0 0 0 0 0 0
Turnip Greens 0 1 0 0 0 0 0









Ra-226 Peaks
_____Pb-214 295 KeV Pb-214 352 KeV Bi-214 609 KeV
Item Net (Counts) Error Net (Counts) Error Net (Counts) Error
Turnip Root and
Green 0 0 0 0 0 0
Turnip Roots 63 33 0 0 0 0
Turnip Roots 0 0 0 0 0 0
Watermelon 0 0 0 0 0 0
Watermelon 0 0 0 0 0 0
Yellow Corn 0 0 0 0 0 0
Yellow Corn 0 0 159 29 126 120
Yellow Corn 0 0 0 0 0 0
Yellow Squash 0 0 0 0 0 0
Yellow Squash 0 0 145 40 183 29
Yellow Squash 35 24 0 0 0 0
Zucchini 32 30 0 0 0 0
Zucchini 0 0 0 0 0 0
Zucchini 0 0 0 0 0 0

Brazil Nuts 4586 167 6936 105 4998 83










Calibration Factors Bkgnd Count Minimum Det. Activity


Pb-210


Ra-226


Pb-210


Ra-226


Pb-210


Ra-226


Item (pCilcps) (pCi/cps) (counts) (counts) (pCi/g) (pCilg)
Beef 1.10E+05 1664 0.000637 0.0062 7.63E-02 3.60E-03
Beef Kidney 1.10E+05 1664 0.000637 0.0062 5.95E-02 2.81 E-03
Black-Eyed Peas 3.40E+05 1670 0.0045 0.001d 4.82E-01 1.41 E-03
Black-Eyed Peas 3.40E+05 1670 0.0045 0.0016 9.10E-01 2.67E-03
Black-Eyed Peas 1.10E+05 1664 0.000637 0.0062 8.98E-02 4.24E-03
Brazil Shells 3.40E+05 1670 0.0045 0.0016 1.02E+00 2.99E-03
Broccoli 1.10E+05 1664 0.000637 0.0062 1.59E-01 7.51E-03
Broccoli 3.40E+05 1670 0.0045 0.0016 1.25E+00 3.66E-03
Broccoli 3.40E+05 1670 0.0045 0.0016 1.08E+00 3.15E-03
Cabbage 3.40E+05 1670 0.0045 0.0016 1.66E+00 4.86E-03
Cabbage 1.10OE+05 1664 0.000637 0.0062 2.37E-01 1.12E-02
Cabbage 1.10E+05 1664 0.000637 0.0062 1.99E-01 9.37E-03
Carrots 3.40E+05 1670 0.0045 0.0016 1.22E+00 3.56E-03
Carrots 1.10E+05 1664 0.000637 0.0062 6.96E-02 3.28E-03
Carrots 1.10OE+05 1664 0.000637 0.0062 1.16E-01 5.49E-03
Cauliflower 1.10OE+05 1664 0.000637 0.0062 1.58E-01 7.46E-0
Cauliflower 3.40E+05 1670 0.0045 0.0016 2.26E+00 6.62E-03
Cauliflower 3.40E+05 1670 0.0045 0.0016 8.82E-01 2.58E-03
Collard Greens 1.10 OE+05 1664 0.000637 0.0062 3.03E-01 1.43E-02
Collard Greens 3.40E+05 1670 0.0045 0.0016 5.71E-01 1.67E-03
Collard Greens 3.40E+05 1670 0.0045 0.0016 2.56E+00 7.49E-03
Corn 3.40E+05 1670 0.0045 0.0016 1.20E+00 3.53E-03
Corn 1.10E+05 1664 0.000637 0.0062 2.13E-01 1.00E-02
Corn 1.10E+05 1664 0.000637 0.0062 1.02E-01 4.80E-03
Cucumber 3.40E+05 1670 0.0045 0.0016 1.00E+00 2.94E-03
Cucumber 3.40E+05 1670 0.0045 0.0016 9.34E-01 2.74E-03
Cucumber 3.40E+05 1670 0.0045 0.0016 1.08E+00 3.16E-03
Eggplant 3.40E+05 1670 0.0045 0.0016 1.46E+00 4.29E-03
Eggplant 1.10OE+05 1664 0.000637 0.0062 1.26E-01 5.97E-03
Eggplant 3.40E+05 1670 0.0045 0.0016 2.68E+00 7.86E-03
Grapefruit 1.10E+05 1664 0.000637 0.0062 9.48E-02 4.48E-03
Grapefruit 1.10E+05 1664 0.000637 0.0062 9.22E-02 4.35E-03
Grapefruit 1.10E+05 1664 0.000637 0.0062 1.01E-01 4.75E-03