gm. 4
, L: a.

.;

 

 

Lt.

.Ll E?

.l yam"??? .r
.

.11....3: .

Mwasmwy .

  

 

x. :v
3.2:.
a.

$5.

 

u. ..
L... .

 

y

11‘3“.

5

 

1 Ar... a
my 1

n..- 4 33......H .lflur...
Jinnah; ‘ ,
1.;

:
w ‘J.

3??

."

 

 

 

 

 

?o|l.§

@§%fi%§§fi3§flfi§

|Il|ll

:9
ham
3 V
L17”. “—

xmu?
xua
.
.:.=
’,a.

“and?

1:31.12: .

 

musu‘4v .‘.,au :
n... 3% X535"

.:u
t
x9

.33..

  

{3005

," Jiffi 65

This is to certify that the
dissertation entitled

A STUDY FOCUSED ON THE RISK OF ILNESS FROM E.
COLI IN RECREATIONAL-USE WATER, USING THE RED
CEDAR WATERSHED AS A MODEL

presented by
MICHAEL J. LANG

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Environmental Toxicology-
Resource Development

 

 

J/pyox/QJé‘ZPW/‘i/
/

Major Professor’s Signature

M47, 1/); 9.4g 4
Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

mt 2153010
U 1 U

41,

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65-p.15

 

 

 

A STUDY FOCUSED ON THE RISK OF ILLNESS FROM E. COLI IN
RECREATIONAL-USE WATER, USING THE RED CEDAR WATERSHED AS A
MODEL
By

Michael J. Lang

A DISSERTATION

Submitted to
7 Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Resource Development

Institute of Environmental Toxicology

2004

ABSTRACT

A STUDY ON THE RISK OF ILLNESS FROM E. COLI IN RECREATIONAL-USE
WATER, USING THE RED CEDAR WATERSHED AS A MODEL

By
Michael J. Lang

With news reports of deaths from the ingestion of E. coli over the past few years,
there is heightened public awareness of the risk of illness fi'om E. coli from swimming
and other recreational uses of surface water. Locally, the State News has run stories on
the high E. coli levels that have been occasionally reported from the weekly sampling of
the Red Cedar on the MSU campus. In addition, high E. coli levels have lead to unsafe
conditions in the Clinton River and Lake St. Clair and cost the State of Michigan $2.5
million in testing to determine the sources of the contamination. Nationwide, a majority
of bodies of water that are used for recreational purposes, are considered to have E. coli
counts above acceptable levels. In developing nations, 90% of untreated sewage from
urban areas is dumped into streams and oceans causing unsafe conditions in both
recreational-use water and drinking water due to high E. coli levels. The sources of E.
coli are thought to be fecal contamination from humans, domestic animals and wildlife, as
well as runoff from agricultural land, inadequate septic systems or sewer overflow.
Runoff from heavy rains may impact E. coli levels in nearby surface water.

In this study, data'collected through a collaboration of State agencies, county
health departments and MSU resources were analyzed and linked to land use and
physical characteristic of the watershed, such as rainfall and nutrient concentrations,

using statistical methods. In addition, a small number of samples from the Red Cedar

were analyzed for both total E. coli and pathogenic E. coli levels using the traditional

methods, as well as, a new innovative method.

Copyright by
Michael J. Lang
2003

ACKNOWLEDGEMENTS

I would like to thank my graduate committee consisting of Drs. Daniel Bronstein,
Scott Witter, Jon Bartholic and John Kaneene for their support and guidance. With out
meetings and discussions with them, this dissertation could not have been completed.

I would like to thank Bob Godbold of Ingham County Health Department, Don
Heyduke of Livingston County Health Department, and Betty Wemette-Babian of the
MSU physician’s office for collecting samples from the Red Cedar River and allowing
me the use of the data generated from the sampling. In addition, I must thank the
Michigan Department of Environmental Quality and the Michigan Department of
Community Health for analyzing samples and communicating data.

I would like to thank Dr. Evangelyn Alocilja, Zarini Muhammad-Tahir, and the
members of the Biosensor Lab in the Agricultural Engineering Department for the use of
the biosensor.

I would like to thank Dr. Sasha Kravchenko for help with the statistical analysis
used in this dissertation. I would like to thank Dr. Don Penner for his help, guidance and
support in finishing this dissertation.

Finally, I must thank my lovely wife Sarah for her support and patience during
my journey through this degree. If I had not been completing this degree, we would not
have met while we were both in graduate school at MSU. This journey has been

intellectually, spiritually, and emotionally fulfilling.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii
LIST OF FIGURES ................................................................................ viii
LIST OF ABBREVIATIONS ........................................................................ x
CHAPTER 1
INTRODUCTION AND OVERVIEW ............................................................ 1
CHAPTER 2

A REVIEW OF THE IMPACT OF E. COLI AND E. COLI 0157:H7 ON HUMAN
HEALTH IN A WATER SHED: RISK FACTORS, RISK ASSESSMENT TOOLS,

EPIDEMIOLOGICAL EVIDENCE, AND PUBLIC POLICY ................................. 8
Escherichia coli ............................................................................ 10
Risk Factors for Illness in Recreation-use Water ...................................... l4
Causal Connection of Illness and Agriculture .......................................... 22
Tools for E. coli Risk Assessment ........................................................ 26
Epidemiological Evidence ................................................................. 33
Public Policy to Reduce Risk in Recreation Water40
Conclusions ................................................................................. 48

CHAPTER 3

THE EFFECT OF RAIN FALL, RIVER FLOW, AND OTHER METEOROLOGICAL

ASPECTS OF E. COLI LEVELS IN THE RED CEDAR RIVER ........................... 73

CHAPTER 4

THE EFFECT OF RAINFALL, NUTRIENT LEVELS, AND LAND-USE ON E. COLI

LEVELS IN THE RED CEDAR RIVER ........................................................ 98

CHAPTER 5

A LINKAGE BETWEEN TECHNOLOGY AND POLICY IN EVALUATING THE

RISK OF MICROBIAL SAFETY I RECREATIONAL-USE WATER .................... 118

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS ............................................ 143

APPENDIX ...................................................................................................................... 148

vi

LIST OF TABLES

Table 1-1. E. coli 015 7:H 7 major outbreaks in the news .................................................. 2

Table 1-2. Collaborators on Research ................................................................................. 6
Table 2-1. Categories of Pathogenic E. coli ..................................................... 11
Table 22. Definition of a Concentrated Animal Feeding Operation (CAFO) .............. 19
Table 2-3. The Nine Minimum Controls for Combined Sewer Overflows .................. 45

Table 3-1 . Number of Weeks that E. coli Concentration Over Full (1000 CFU/ml) or
Partial (300 CFU/ml) Body Contact .............................................................. 87

Table 3-2. Statistical Analysis of E. coli Concentration vs. River Flow, Rainfall, Air
Temperature, Humidity, Soil Moisture, Ground Temperature, Solar Radiation, Wind
Speed, Water Temperature, and Duck Count. (11 = 96) ........................................ 88

Table 4-1. Statistical Analysis of E. coli Concentration vs. Rainfall, Ammonia Nitrate,
Total Phosphorus, and Total Suspended Solids (11 = 221) .................................... 109

Table 5-1. E. coli Concentration, Rainfall, River Flow, and E. coli 0157:H7
Concentration from the Sampling Sites in the Red Cedar River Watershed ............... 132

Table 5-2. Statistical Analysis of E. coli Concentration vs. Rainfall, River Flow, and
E. coli 0157:H7 Concentration .................................................................. 134

Table 5-3.E. coli 015 7:H7 Analyzed by Traditional Methods and by the Biosensor ..... 137

Table A-1 . Farm Lane Three Years of Data .................................................................... 149
Table A-2. Yearly Mean of E. coli Concentration, Flow and Rainfall ............................ 153
Table A-3. Nutrients, E. coli, and Land-use Data ............................................................ 154

vii

LIST OF FIGURES

Table 1-1. E. coli 015 7:H7 major outbreaks in the news ................................................... 1
Table 1-2. Collaborators on Research .................................................................................. 6
Figure 2-1. Risk Assessment Model .............................................................. 31
Figure 3-1. View of the Farm Lane Bride over the Red Cedar River from the USGS
Gauging Station on the Campus of Michigan State University ............................... 77
Figure 3-2. Locations of where measurements were taken for this study ................... 78

Figure 3-3. The Red Cedar River Watershed with The Michigan State University campus
outlined ............................................................................................... 79

Figure 3-4. E. coli Concentration from April to November in the Red Cedar River. . .....81

Figure 3-5. Flow Rate from April to November in the Red Cedar River .................... 82
Figure 3-6. Mean Values of the Log-transformed E. coli Concentrations in Spring,
Summer and Fall of the Studied Years ........................................................... 83
Figure 3-7. Mean Values of the Log-transformed Flow Rates in Spring, Summer and Fall
of the Studied Years ................................................................................ 83
Figure 3-8. Yearly Mean and Standard Deviation of E. coli Levels and Flow ............. 85
Figure 3-9. Yearly Mean of E. coli Levels, Flow, and Rainfall .............................. 86
Figure 3-10. Probability of E. coli >300 cfm/ 100ml Resulting from Logistic Regression
Based on Cumulative 72 Hour Rainfall .......................................................... 91
Figure 3-11. Probability of E. coli >300 cfm/ 100ml Resulting from Logistic Regression
Based on Cumulative 72 Hour Rainfall and Low Air Temperature .......................... 92
Figure 3-12. Probability of E. coli >1000 cfm/ 100ml Resulting from Logistic Regression
Based on Cumulative 72 Hour Rainfall .......................................................... 94
Figure 4-1. The Red Cedar River Watershed with The Michigan State University campus
outlined ............................................................................................. 102
Figure 4-2. Means E. coli vs. Land use for 17 Selected Sampling Points .................. 105
Figure 4-3. Mean Ammonia Nitrate vs. Land use for 17 Select Sampling Points. 106

viii

LIST OF FIGURES

Figure 4-4. Mean Total Phosphorus vs. Land use for 17 Select Sampling Points... ......107

Figure 4-5. Mean Total Suspended Solids vs. Land use for 17 Select Sampling Points.108

Figure 4-6. Mean E. coli Levels vs. Land use for all 38 Sampling Points... . . . . . . . ........1 11
Figure 5-1. Red Cedar River Watershed with Locations of Sampling Sites ............... 125
Figure 5-2. Schematic diagram of the immunosensor .............. ' ........................... 129

Figure 5-3. Resistance drop of the biosensor tested in distilled water samples inoculated
with E. coli 0157:H7 .............................................................................. 136

ix

AF 0
AGNPS
AU
BEACH
BMP
CAFO
CFU
CSO
CWA
DAEC
DMF

E. coli
EAEC
EHEC
EIEC
EPA
EPEC
ETEC

GIS
HUS
ICHD
IMS
MDCH
MDEQ

MSU
NC
NLCD
NPDES
NRDC
NWS
PBS
PCR
PF GE
POTW
SSO
TMDL
USDA
USEPA
USPHS
USGS
WHO
WWTP

LIST OF ABBREVIATIONS

Animal Feeding Operations

Agricultural Non-Point Source

Animal Units

Beach Environmental Assessment Closure and Health Act
best management practices

Concentrated Animal Feeding Operation
Colony Forming Unit

Combined Sewer Overflow

Clean Water Act

Diffusely Adherent E. coli

N, N Dimethylformamide

Escherichia coli

Enteroaggregative E. coli
Enterohemorrhagic E. coli

Enteroinvasive E. coli

Environmental Protection Agency
Enteropathogenic E. coli

Enterotoxigenic E. coli

grams

Global lnforrnation Systems

Hemolytic Uremic Syndrome

Ingham County Health Department
Immunomagnetic separation

Michigan Department of Community Health
Michigan Department of Environmental Quality
milliliters

Michigan State University

Nitrocellulose

National Land Cover Data

National Pollution Discharge Elimination System
National Resources Defense Council
National Weather Service

phosphate buffer saline

Polymerase Chain Reaction

Pulsed-field gel electrophoresis

Publicly Owned Treatment Works

Sanitary Sewer Overflows

Total Maximum Daily Load

United States Department of Agriculture
United States Environmental Protection Agency
United States Public Health Services

United States Geological Services

World Health Organization

Waste Water Treatment Plant

Email

Introduction and Overview

Introduction

With news reports of deaths from the ingestion of E. coli over the past few years,
there is heightened public awareness of the risk of illness from E. coli from swimming
and other recreational uses of surface water. Locally, the State News has run stories on
the high E. coli levels that have been occasionally reported from the weekly sampling of
the Red Cedar on the MSU campus (Johnson 2003; Cynecki 2003). In addition, high E.
coli levels have lead to unsafe conditions in the Clinton River and Lake St. Clair and cost
the State of Michigan $2.5 million in testing to determine the sources of the
contamination (Morris 2002). Nationwide, a majority of bodies of water that are used for
recreational purposes, are considered to have E. coli counts above acceptable levels
(Dorfman 2002). In developing nations, 90% of untreated sewage from urban areas is
drunped into streams and oceans causing unsafe conditions in both recreational-use water
and drinking water due to high E. coli levels (Crossette 1996). The sources of E. coli are
thought to be fecal contamination from humans, domestic animals and wildlife, as well as
runoff from agricultural land, inadequate septic systems or sewer overflow (U SEPA 1986
and 2000). Runoff from heavy rains may impact E. coli levels in nearby surface water.

A causal connection between fecal coliforrn and gastrointestinal sickness was first
identified in 1953. (Stevenson 1953). In the late 19708, in a landmark prospective cohort
study, reported a linear relationship between the incidence of gastroenteritis among

swimmers and marine bacterial counts (Cabelli 1982). There has a large body of

literature involving the epidemiology of illness from the use of recreational water. An
excellent review of this body of literature has been published which included 37 studies.
This review concluded that a causal dose-related relationship between gastrointestinal
symptoms and recreational water quality measured by bacterial indicator counts exists
(Pruss 1998).

E. coli 0157 was first reported in the United States in 1982, when it was
associated with a multi-state outbreak of hemorrhagic colitis that was traced to
hamburger from a restaurant chain (Riley 1983; Tarr 1995). Exposure through ingestion
of E. coli 015 7:H7 can cause severe bloody diarrhea and abdominal cramps although in
some cases the infection causes non-bloody diarrhea. Usually little or no fever is present,
and the illness resolves in 5 to 10 days. The clinical diagnosis of this illness is
haemorrhagic colitis. In some people, particularly children under 5 years of age and the
elderly, the infection can also cause a complication called hemolytic uremic syndrome
(HUS), in which the red blood cells are destroyed and the kidneys fail (Moake 1994;

CDC 2001).

Table 1-1. E. coli 015 7:H7 major outbreaks in the news.

Timeframe Location Company Vector for Infection
1982 Washington State Jack in the Box Hamburger

1982 Michigan Small cider mills Cider

June 1997 Nationwide Hudson Meats Hamburger

May 2000 Walkerton, Ont. Municipal water Drinking water
August 2000 Washington, NY State Fair Drinking water
July 2002 Nationwide ConAgra Hamburger

The original intent of the research was to perform a risk assessment for illness
from E. coli 015 7:H7 to humans when using the Red Cedar River for recreational
purposes. Since there are no dose response data for illness caused by E. coli 015 7:H7 in
recreational use water or even from ingestion via contaminated food, epidemiological
studies on illness from E. coli 015 7:H7 in recreational use water were to be collected and
meta-analysis used to determine the dose response for E. coli 015 7:H7 in recreational
use water. Unfortunately, there is not enough data from the few epidemiological studies
that have been reported (see Chapter 2). Since the data that was collected in this study
was E. coli concentration levels, a relationship between E. coli concentration and E. coli
015 7:H7 concentration was necessary. In an extensive literture search, no such
relationship has been reported or estimated. Since the original goals of the research were
not possible, the research questions were changed so that some of the identified holes in
the reported literature could possibly be filled by this study. Results from this study may
provide mechanisms for improved identification of potential risk factors for illness from
E. coli in recreational-use water and may provide a solution for the real-time

communication of such risks.

Research Questions

1. Are there physical measurements that are taken in a watershed that maybe indicators
of an increased risk to humans using recreational water from exposure form E. coli?
2. Does of the type of land use surrounding the sampling point impact change in the

reported physical measurement that are taken in the Red Cedar River watershed?

3. Do the levels of E. coli 0157:H7 in recreational water in the Red Cedar River follow
the levels of the indicator organism that is monitored and regulated?
4. What is the risk to humans using in recreational water from pathogenic E. coli

0157:H7 in the Red Cedar River?

Overview

In this study, data collected through a collaboration of State agencies, county
health departments and MSU resources were analyzed and linked to land use and
physical characteristic of the watershed, such as rainfall and nutrient concentrations,
using statistical methods. In addition, a small number of samples from the Red Cedar
were analyzed for both total E. coli and pathogenic E. coli levels using the traditional
methods, as well as, a new innovative method. This dissertation is set up so that chapters
2 through 5 are stand-alone manuscript written for submission to appropriate journals.

Chapter 2 entitled “A Review of the Impact of E. coli and E. coli 015 7:H7 on
Human Health in a Watershed: Risk Factors, Risk Assessment Tools, Epidemiological
Evidence, and Public Policy” is an extensive literature review of E. coli contamination in
recreational use water and the effect on human health. Many of the references in this
review are also used in later chapters. This chapter is written with the intent to submit it
as a manuscript to Epidemiological Reviews.

Chapter 3 entitled “The Effect of Rainfall, River Flow and other Meteorological
Aspects on E. coli Levels in the Red Cedar River” has the following objectives. An
objective of this study is to determine if there is a seasonal effect on the concentration

levels of E. coli in the Red Cedar River watershed. Another objective is to determine if

there is relationship between the flow rate of the river and concentration levels of E. coli.
In addition, an objective of this study is to determine the statistical significance of
concentration levels of E. coli as related to the watershed physical characteristics of
rainfall, air temperature, water temperature, intensity of the sun, humidity, soil moisture,
ground temperature, wind speed, duck population, as well as river flow. Finally, an
objective of this study is to use the statistical data to model risk factors to so that water
safety can be determined by the changes in physical characteristics which maybe
determinable in real time. It is hypothesized, based on the results reported in the
literature, that the model risk factors are directly related to rainfall and associated factors.
This chapter is written with the intent to submit it as a manuscript to Journal of
Environmental Quality.

Chapter 4 entitled “The Effect of Rainfall, Nutrient Levels, and Land-use on E.
coli Levels in the Red Cedar River” has the following objectives. The objective of this
study is determine if the type of land-use around a sampling point is associated with
elevated E. coli concentration levels. Another objective of this study is to determine if
the type of land-use around a sampling point is associated with elevated levels of
nutrients. In addition, an objective of this study is determine if there is any association
between the concentration levels of nutrients and the concentration levels of E. coli. It is
hypothesized that agricultural land-use has the highest discharges of E. coli and nutrients
and therefore has a greater negative impact to the Red Cedar River than other types of
land-use. This chapter is written with the intent to submit it as a manuscript to Journal of

American Water Resources.

Chapter 5 entitled “A Linkage between Technology and Policy in Evaluating the
Risk of Microbial Safety in Recreational-use Water” has the following objectives. The
main focus of this project is to improve the safety of recreational-use water through better
communication of the actual risk. An objective is to determine if rain or river flow has an
influence on the concentration of E. coli and E. coli 0157:H7. Another objective is to
evaluate if total E. coli measured maybe used as an indicator of pathogenic E. coli
0157:H7 contamination. In addition, an objective is to evaluate if a biosensor maybe
employed to measure quickly the presence of E. coli and E. coli 0157:H7. It is
hypothesized that a linear relationship will exist between E. coli concentrations and E.
coli 015 7:H7 concentrations. This chapter is written with the intent to submit it as a

manuscript to Environmental Health Perspectives

Table 1.2. Collaborators on Research.

-Ingham County Health Department

oLivingston County Health Department
~Michigan Department of Environmental Quality
-Michigan Department of Community Health
oMSU Physician’s office

OMSU Biosensor Lab

-MSU WATER Human Health Subcommittee
oMSU Department of Crop and Soil Sciences
0East Lansing WWTP

References

Cabelli VJ, Dufour AP, McCabe LJ, Levin MA. 1982. Swimming-associated
gastroenteritis and water quality. Am J Epidemiol 115:606-616.

Cabelli VJ. 1989. Swimming-associated illness and recreational water-quality criteria.
Water Science and Technology 21(2): 13-21.

CDC. 2001 . Escherichia col i 015 7:11 7.

http://wwwcdc. gov/11cidod/dbmd/tliscaseinlb/esclreriChiacoli_g.htm

Crossette B. 1996. Hope and pragmatism, for UN. cities conferences. New York Times,
3 June A3.

Cynecki K. 2003. Area river safe after sewage leaks. The State News. 25 Aug.

Dorfman M. 2002. Testing the waters XII: A guide to water quality at vacation beaches.
NRDC, 211pp.

Johnson A. 2002. Ingham county seeks source for bacteria in Red Cedar River. The State
News. 1 Apr.

Moake J L. 1994. Hemolytic Uremic Syndrome: Basis Science. Lancet 343:393-397.
Morris 1. 2002. Water tests to cost $2.5M. The Detroit News. 4 Aug. BS.

Pruss A. 1998. Review of epidemiological studies on health effects from exposure to
recreational water. Int J Epidemiol. 27(1): 1-9.

Riley LW, Remis RS, Helgerson SD. 1983. Hemorrhagic colitis associated with a rare
Escherichia coli serotype. N Engl J Med;308:681-685.

Stevenson AJ. 1953. Studies of bathing water quality and health. American Journal of
Public Health 43:529-53 8.

Tarr PI. 1995. Escherichia-Cali 015 7-H7 - Clinical, Diagnostic, and Epidemiologic
Aspects of Human Infection. Clinical Infectious Diseases 20(1): 1-10.

USEPA Human Health Risk Assessment Guidelines (1986), and supplement (August
2000)

Cha ter 2
A Review of the Impact of E. coli and E. coli 015 7:11 7 on Human Health in a

Watershed: Risk Factors, Risk Assessment Tools, Epidemiological Evidence, and

Public Policy

Table of Contents

Introduction

Escherichia coli

Risk Factors for Illness in Recreation-use Water

Physical Characteristics of a Watershed
Agricultural Animals and Activities
Other Animals Reservoirs

Causal Connection of Illness and Agriculture

Human Sewage
Causal Connection of Illness and Sewage

Tools for E. coli Risk Assessment
Modeling of the Watershed

Detection Methods for E. coli
Risk Assessment Models
Epidemiological Evidence
Indicator Bacteria
E. coli 015 7:H7
Public Policy

Conclusions

References

Introduction

Beach closings and illness after exposure to marine water may be increasing in
frequency (Harvell 1999; Harvard Med School 1998)). In the United States from 1988 to
1994, there were over 12,000 coastal beach closings and advisories (an increase of 400%
over that period), with over 75% of the closings due to microbial contamination (Barton
1995). A recent report compiled by the Natural Resources Defense Council (NRDC),
which surveyed more than 200 waterfront communities, found that during 1999 there
were at least 6,160 beach days of closings and advisories at beaches (Dorfman 2000).
According to NRDC’s twelfth annual beach report, at least 13,410 closings and advisories
were issued across the country in 2001, a 19 percent jump over the previous year
(Dorfman 2002). Although not all cases can be traced to anthropogenic discharges, the
belief that marine pollution has no significant impact on health (Carson 1951; Moore
1959) must now be challenged. Infectious diseases and toxin-related illness may be
caused by enteric pathogens or chemicals that enter the marine environment from
terrestrial ecosystems (e.g., through fecal contamination from a number of point and
diffuse sources) (Weiskel 1996). Alternatively, indigenous organisms and their associated
biotoxins that may have increased in number or vinrlence as a result of ecological
imbalance (Epstein 1993; Smayda 1993) can cause negative health effects.
Anthropogenic inputs to the coastal environment may be contributing to both terrestrial

and marine stress (Epstein 1993).

There is a large body of literature involving the illness caused by E. coli strains in

water. In the earlyl9803, E. coli 015 7:H7 was recognized as a pathogen (Riley 1982).

The pathogenicity to humans from this pathogen has been reported to be as low as 10
cells (Phillips CA 1999). There are an estimated 73,000 cases of E. coli 0157 infections
per year in the United States, of which approximately 62,000 are food-bome and 11,000
are waterborne (Mead PS et a1. 1999). These estimations consider waterborne cases from
ingestion of water as food not from recreational use of a waterbody. Examples of
waterborne cases include Walkerton, Ontario, where the town’s water supply was
contaminated by E. coli 015 7:H7 (not included in the estimations because it happened in
Canada in the year 2000) (N ikiforuk 2000; Brooke 2000; Wickens 2000) and Washington
County fair in the state of New York (Bopp 2003). Although there are an estimated
73,000 cases of E. coli 0157 infections per year in the United States since 1982 (CDC

2000), reports of clinical cases from E. coli 015 7:H7 are less than ten worldwide.

Escherichia coli

Escherichia coli (E. coli) is the type species of the genus Escherichia, which
contains mostly motile rod-like gram-negative bacilli within the family
Enterobacteriaceae and the tribe Escherichia. E. coli bacteria live in the digestive
systems of humans and other warm-blooded animals. E. coli can be found in the fecal
flora of a wide variety of animals including cattle, Sheep, goats, pigs, cats, dogs, chickens,
and gulls (Hancock 2001; Niemis 1991).

The organism typically colonizes the infant gastrointestinal tract within hours of
life, and, thereafter, E. coli and the host derive mutual benefit (Drasser 1974). E. coli
usually remains harmlessly confined to the intestinal lumen; however, in the debilitated or

immunosuppressed host, or when gastrointestinal barriers are violated, even normal

10

"nonpathogenic" strains of E. coli can cause infection. Most strains of the E. coli bacteria
are not dangerous, but they can indicate the presence of other disease-causing bacteria.

There are a variety of sources that contribute bacteria and other pathogens to the
surface water. These sources include illicit waste connections to storm sewers or roadside
ditches, septic systems, combined and sanitary sewer overflows, storm (rain) runoff, wild
or domestic animal waste, and agriculture runoff.

E. coli 0157 was first reported in the United States in 1982, when it was
associated with a multi-state outbreak of hemorrhagic colitis that was traced to
hamburger from a restaurant chain (Riley 1982; Tarr 1985). The current teachings

suggest that there are Six categories of pathogenic E. coli as described in Table 2-1.

Table 2-1. Categories of Pathogenic E. coli (adapted from Nataro and Kater 1998)

 

Examples of
cronym Name Associated Clinical Syndromes Serotypes

 

ETEC Enterotoxigenic E. coli ifraveler's diarrhea

 

EPEC Enteropathogenic E. coli Watery diarrhea of infants

 

Hemolytic-uremic syndrome

 

 

EHEC Enterohemorrhagic E. coli (HUS) 015 7:H7
EAEC Enteroaggregative E. coli Persistent diarrhea 0127:H2
EIEC Enteroinvasive E. coli

 

 

 

 

 

DAEC Diffusely Adherent E. coli

 

 

Of these Six, two categories, EIEC and DAEC, are considered “emerging
pathogens.” There is very little documentation on these categories and the identification
of these strains has been limited to infants in less developed countries. In addition, these
emerging strains have not been identified in the developed countries. For the other

categories, three of these, ETEC, EPEC, and EAEC, have doses that cause illness in

humans that are far above the standards of total E. coli that is acceptable for “safe”
human recreational use of surface water. If the waterbody has levels of E. coli below the
regulated standard then there is no risk from these three categories of pathogenic E. coli.
The final category of pathogenic is EHEC which includes the famous serotype 0] 5 7H 7.
The infectious dose for this strain has been estimated in the literature to be anywhere
from 1000 organisms to just 10 organisms (Philips 1999). These levels are far below the
total E. coli standard that is acceptable for “safe” human recreational use of surface
water. No literature references are available for what the percentage of EHEC in a total
E. coli measurement from surface water may be, therefore, a huge gap in the literature
exists and research is necessary to help answer these questions.

Among the most important virulence characteristics of E. coli 0157 is its ability
to produce one or more Shiga toxins also called verocytotoxins and formerly known as
Shiga-like toxins. The first of these, Shiga toxin 1 (Stxl) is indistinguishable from Shiga
toxin produced by Shigella dysenteriae type 1 (Nataro and Kaper 1998). The second,
Shiga toxin 2 (Stx2) is a more divergent molecule with only 56% amino acid homology
with Stxl . Most E. coli 0157 strains produce Stx2 with the percentage that also produce
Stxl ranges from less than 25% in series from Europe (Thomas 1996) to greater than
80% in a series from North America (Cimolai 1994; Slusker 1997) and Japan (NIH
1996)

Shiga toxin infection often causes severe bloody diarrhea and abdominal cramps
although in some cases the infection causes non-bloody diarrhea. Usually little or no
fever is present, and the illness resolves in 5 to 10 days. The clinical diagnosis of this

illness is haemorrhagic colitis. Ground beef and other bovine products have often been

12

implicated as sources, along with other food products (Ackers 1998, Besset 1993; Hilbom
1999; Michino 1999; Tamblyn 1999; Watanabe 1996) and person-to-person transmission
(Pavia 1990; Bender 1997). Occasional outbreaks have also been associated with public
drinking water (Swerdlow 1992) and swimming in contaminated water (Friedman 1999).

In some people, particularly children under 5 years of age and the elderly, the
infection can also cause a complication called hemolytic uremic syndrome (HUS), in
which the red blood cells are destroyed and the kidneys fail (Riley 1983; Wells 1983;
Moake 1994). About 2%-7% of infections lead to this complication. In the United States,
HUS is the principal cause of acute kidney failure in children, and most cases of HUS are
caused by E. coli 01 5 7:H 7. HUS is a life-threatening condition usually treated in an
intensive care unit. Blood transfusions and kidney dialysis are often required. With
intensive care, the death rate for HUS is 3%-5% (CDC 2001). About one-third of
persons with HUS have abnormal kidney function many years later, and a few require
long-term dialysis. Another 8% of persons with HUS have other lifelong complications,
such as high blood pressure, seizures, blindness, paralysis, and the effects of having part
of their bowel removed (CDC 2001).

Human infection with E. coli 0157 has been reported from over 30 countries on
six continents. Annual incidence rates of 8 per 100,000 population or greater have been
reported in regions of Scotland (Rielly 1997; Coia 1998; Licence 2001), Canada (Waters
1994), and the US. (Griffin 1991). High rates may be present in regions of South
America, especially in Argentina where HUS has an incidence 5-10 times higher than
North America (Lopez 1989; Rivas 1998). Infection rates appear to be rather low in

Australia (Fegan 2002).

I3

Risk Factors for Illness in Recreation-use Water

Ever Since fecal contamination of water was determined a human health risk, there
has always been a great deal of concern regarding the level of coliforrn bacteria counts in
water. Many bodies of water throughout the world are considered to have counts above
acceptable levels. The sources of these coliforms are thought to be fecal contamination
from humans, domestic animals and wildlife, as well as runoff from agricultural land,

inadequate septic systems or sewer overflow (EPA 2000).

Physical Characteristics of a Watershed

Physical characteristics of the watershed are potential risk factors to the levels of
E. coli and its various strains. The physical factors that will be reviewed in this section
include seasonal variability, rainfall, river flow, nutrient, and the rate of survival of
bacteria in water.

Over a twelve-year period (1984-1995) a study examined admissions for HUS in
Scotland and determined that seasonality was present with the highest amount of
admissions in July and August (Douglas 1997). It should be noted that this study was not
limited to recreational use of water and showed the seasonal pattern only for patients
under 15 years old. In contrast, a study in Florida showed that concentration of fecal
indicator organisms during the late fall and early winter months which corresponds to the
wet-weather months in Florida (Lipp 2001). In addition the study determined that the
levels of fecal indicators were significantly associated with rainfall, stream flow, and
temperature. Moreover, a study of two stearns in Arkansas over a three year period

concluded that concentrations of indicator bacteria increased with increasing flow rates

and seasonal effects were observed on the indicators with the highest levels occurring
during the summer months (Edwards 1997). Furthermore in a study of a dairy herd
reported that E. coli 015 7:H7 was found in 4.3% of the herd and peaked during the May
to July timefiame but was not found in the herd from November to May (Mechie 1997).
However a study in northwest England reported that bacteria indicators Showed no
seasonal variations over a two-year period (Obiri-Danso 1999).

Other studies support the association of the levels of fecal indicators in a water
body and rainfall events. In a study that examined 99 samples from three tributaries that
contributed to different drinking water reservoirs showed that E. coli levels along with
other bacteriological parameters increased considerably during extreme rainfall events
(Kistemann 2002). In a related study on Delaware River, increased concentrations of
Giardia, Cryptosporidium and other microorganisms were associated with rainfall
(Atherholt 1998). In a prospective study on waterborne disease outbreaks inthe US. for
1948 to 1994, analyzed 548 reported outbreaks as documented by the USEPA database.
This results from this study showed that 51% of waterborne disease outbreaks were
preceded by precipitation events above the 90th percentile and 67% by events above the
80th percentile (Curriero 2001). This study concluded that there is a statistical
significance association between rainfall and waterborne disease outbreaks. Other
studies also conclude that E. coli levels increase with rainfall (Briski 2000; Ferguson
1996; Atherholt 1998; Pettibone 1996; Niemi 1991; Mallin 2001). A recent study
suggests that an increase in rainfall or snovvrnelt increases the impact of diseases caused

by microbiologic agents (Rose 2001).

IS

Increased river flow has been related to increases in the levels of E. coli in the
river. Early studies in Idaho and Oregon found that fecal coli concentrations were higher
during periods of high flow than during period of lower flow (Stephenson 1978;
Tiedemann 1988). These finding were confirmed in a study of two rivers in Arkansas
which concluded that concentrations of indicator bacteria increased with increasing flow
rates (Edwards 1997).

As far as nutrients and their relation to E. coli there has not been many reports in
the literature. Total suspend solid were strongly correlated to indicator bacteria when
flow rates were the highest (Pettibone 1996). In another study, a statistical analysis
showed that there was a significant correlations between levels of nitrate as well as levels
of phosphate with the indicator bacteria faecal coliforrn (Baby 2002).

In a recent study, model systems were used to determine the persistence of E. coli
0157 in river water, cattle faeces, and soil cores (Maule 2000). Survival of the organism
was found to be greatest in soil cores containing rooted grass. Under these conditions
viable numbers were shown to decline from approximately 108 /g soil to between 106 and
107 /g soil after 130 days. When the organism was inoculated into cattle faeces it
remained detectable at high levels for more than 50 days. In contrast the organism
survived much less readily in cattle slurry and river water where it fell in numbers fi'om

more than 106/ml to undetectable levels in 10 and 27 days, respectively (Maule 2000).

The survival characteristics of a mixture of five nalidixic acid-resistant E. coli
0157.'H 7 strains innoculated at 103 CFU/ml in filtered and autoclaved municipal water, in
reservoir water, and in water from two recreational lakes were determined for a period of

91 days and stored at three different temperatures of 8, 15, or 25 degrees C. Greatest

survival was in filtered autoclaved municipal water and least in lake water. Regardless of
the water source, survival was greatest at 8 degrees C and least at 25 degrees (Wang
1998)

In study E. coli 0157 was investigated to determine if it would multiply in a
medium containing 5% NaCl and in sterilized marine water. Results indicate that E. coli
0157 could survive in unsterilized marine water for at least 15 days. On the basis of these
results, it was postulated that E. coli 0157 may survive in natmal marine water (Miyagi
2002)

Environmental survival of Escherichia coli 0157 may play an important role in
the persistence and dissemination of this organism on farms. The survival of culturable
and infectious E, coli 0157 was studied using microcosms Simulating cattle mater
troughs. Culturable E, coli 0157 survived for at least 245 days in the microcosm
sediments (LeJeune 2001). In a study aimed to investigate the survival characteristics of
Escherichia coli 015 7:H 7 in farm water, and in sterile distilled municipal water reported
that the organism survived in farm water for over 31 days and in the distilled water for 17
days (McGee 2002).

The survival characteristics of Escherichia coli 015 7:H7 in private drinking
water wells were investigated to assess the potential for human exposure. A non-
toxigenic, chromosomally lux-marked strain of E coli 01 5 7.1-] 7 was inoculated into well
water from four different sites in the North East of Scotland. These waters differed
significantly in their heavy metal contents as well as nutrient and bacterial grazer

concentrations. Grazing and other biological factors were studied using filtered (3 and 0.2

17

mum) and autoclaved water. The survival of E. coli 015 7:H7 was primarily decreased by
elevated copper concentrations (Artz 2002)

Escherichia coli 015 7:H 7 was inoculated at final concentrations of 103 or 106 /ml
into natural non-carbonated mineral water (MW), sterile natural mineral water (SMW)
and sterile distilled deionized water (SDDW) and stored at 15 degrees C for 10 weeks.
Samples were examined every 7 d for the presence of E. coli 01 5 7:H 7. There was a
Significant difference in the survival of E. coli 0157:H7 (103 /ml inoculum) between the
MW and the SDDW and between the MW and the SMW with the pathogen surviving
longest in the MW samples. In contrast, at 106 /ml, no significant differences in the
survival of E. coli 015 7:H7 were observed between the water types (Kerr M 1999).

In addition studies have reported that E. coli survives and even grows in
freshwater and marine sediment (Gerba 1976; Hood 1982; LaLiberte 1982). Sediments
may contain 100 to 1000 times as many fecal indicator bacteria as the overlaying water
(Ashbolt 1993; Van Donsel 1971). E. coli survived over 28 and as long as 68 days in

sediment (Davies 1995).

Agg'cultural Animals and Activities

Animal Feeding Operations (AF Os) are agricultural operations where animals are
kept and raised in confined situations. AF Os generally congregate animals, feed, manure,
and production operations on a small land area. Feed is brought to the animals rather than
the animals grazing or otherwise seeking feed in pastures. Animal waste and wastewater
can enter water bodies from spills or breaks of waste storage structures (due to accidents

or excessive rain), and non-agricultural application of manure to cropland. AFOS that

18

meet the regulatory definition of a Concentrated Animal Feeding Operation (CAFO) have
the potential of being regulated under the NPDES permitting program. A facility is an
AF 0 if animals are stabled/confined, or fed/maintained, for 45 days or more within any
12-month period, and the facility does not produce any crops, vegetation or forage

growth (40 CFR 122.230))(1).

Table 2-2. Definition of a Concentrated Animal Feeding Operation (CAFO)
A CAFO is an AFO which:
. Has more than 1,000 animal units (AU), or
. Has 301 to 1,000 AU and wastes are discharged through man-made conveyance

or directly into US waters, or
. Is designated a CAFO by the permitting authority on a case-by-case basis

Animal units (“AUS”) are defined in EPA'S current regulations at 40 CFR 122 and vary
by animal type. An AU is equivalent to one beef cow or 2.5 mature swine or up to 100
chickens. Winter-feeding of animals on pasture or rangeland is not normally considered
an AF 0.

USDA reports that there were 1.2 million livestock and poultry operations in the
United States in 1997 (USDA 1999). This number includes all operations that raise beef
or dairy cattle, hogs, chickens (broilers or layers), and turkeys, and includes both
confinement and non-confinement (i.e., grazing and rangefed) production. Of these, EPA
estimates that there are about 376,000 AF Os that raise or house animals in confinement,
as defined by the existing regulations.

AF Os (including CAFOS) produce and manage large amounts of animal waste,
most in the form of manure. USDA estimates that 291 billion pounds (132 million metric

tons) of "as excreted" manure were generated in 1997 from major livestock and poultry

l9

operations (USDA 1999). The scale of this unprecedented outpouring of animal waste is
staggering: 130 times the waste generated by humans in this country each year (US
Senate Comm. Report 1997).

Recent trends across the US. livestock and poultry sectors are marked by a
decline in the number of operations due to ongoing consolidation in the animal
production industry (MacDonald 2000; McBride 1997). Increasingly, larger, more
industrialized, and highly specialized operations now account for a greater share of all
animal production. This concentrates more animals, and thus more manure and
wastewater, in a single location, and raises the potential for significant environmental
damages unless manure is properly handled. USDA reports that there were 1.1 million
livestock and poultry farms in the United States in 1997, about 50 percent fewer than the
2.3 million farms reported in 1974 (USDA 1999 and 1976). Since the 1970s, the
combined forces of population growth and re-location of operations closer to consumer
markets and processing sectors have resulted in more AF Os located near densely
populated areas. Surface waters in these areas face additional stresses from urban runoff
and other point sources. The proximity of large AF Os to human populations thus
increases the potential for human health impacts and ecological damage if manure and
wastewater at AF Os is improperly discharged.

The most important animal species in terms of as a vextor for human infection is
cattle. High rates of colonization of six-positive E. coli have been found in bovine herds
in many countries (Burens 1995; Clarke 1994; Griffin 1991; Hancock 1994; Wells 1991).
These rates are as high as 60% but are more typically in the range of 10 to 25%. Stx-

producing E. coli strains are usually isolated fi-om healthy animals but may be associated

20

with an initial episode of diarrhea in young animals followed by asymptomatic
colonization. The isolation rates of 015 7:H7 are much lower than those of non-015 7:H7
serotypes. Surveys of US. dairy and beef cattle have found E. coli 015 7:H7 in 0 to 2.8%
of animals, with the highest isolation rates reported from younger rather than older
animals (Hancock 1994; Wells 1991). In 1986, E. coli 0157 was recovered from healthy
dairy cows, suggesting that dairy and beef herds could serve as a reservoir (Martin 1986).
Subsequent studies have confirmed that E. coli 0157 and other EHEC strains are
commonly found in beef and dairy cattle (Martin 1986) as well as animals associated with
farm as well as sheep, pigs, goats and chickens (Ogden 2002; Kariuki 1999; Griffin 1991;
Strachan 2001).

Environmental survival of Escherichia coli 0157 may play an important role in
the persistence and dissemination of this organism on farms. The survival of culturable
and infectious E. coli 0157 was studied using microcosms Simulating cattle mater
troughs. E. coli 0157 strains surviving more than 6 months in contaminated microcosms
were infectious to a group of 10-week-old calves (LeJeune 2001). Fecal excretion of E.
coli 0157 by these calves persisted for 87 days after challenge. Water trough sediments
contaminated with feces from cattle excreting E. coli 0157 may serve as a long-term
reservoir of this organism on farms and a source of infection for cattle.

E. coli has been identified in cattle on all continents with the exception of
Antarctica. In Asia, the pathogen has been identified in Chinese and Japanese cattle
(Zhou 2002; Elliot 200]; Dundas 2000; Tanaka 2000) In Europe E. coli 015 7 has found
in cattle in the Netherlands, Norway, Finland, Britian, Germany, France, Chezsolvokia,

(Beutin 2000; Chalmers 2000; Geue 2002; Heuvelink 1999; Heuvelink 1998; Jones 1999;

21

Lahti 2001; Osek 1999; Osek 2002; Vermozy-Rozand 2002; Vold 2001). In North
America, E. coli 0157 has been identified in cattle and other farm animals in the United
States and Canada, as well as South America (Bielaszewska 2000; Hancock 2001; Hoar
2001; Joseph 2002; Laegreid 1999; License 2001; Mariani-Karkdjian 1999; McClure
2000; Mead 1998; Padhye 1992; Philips 1999; Notario 2000; Rasmussen 2001;
Samadpour 2002; Sargeant 2000; Verweyen 2000; Wang 1996; Whipp 1994). E. coli

015 7:H 7 is least prevalent in Australia (Elliot 2001; Fegan 2002).

Other Anirn:al Reservoirs

Some studies suggest that elk, deer, ducks, birds and other wild animals may also
be a source of E. coli 0157 (Niemi 1991; Wallace 1997; F eldman 2002; Olson 2002;
Samadpour 2002; Rice 2003). Companion animals including cats, dogs, reptiles, and
birds have been reported to have E. coli and/or E. coli 015 7:H7 in their feces (Beutin
1993; Wallace 1997; Synge 2000; Hancock 2000; Enriquez 2001). In addition fecal
samples from 300 zoo animals have been analyzed and only six animals, a horse and five

species of primates, were positive for E. coli 015 7:H7 (Bauwens 2000).

Causal Connection of Illness and Agriculture

In 1999, an E. coli outbreak occurred at the Washington County Fair in New York
State. This outbreak was possibly the largest waterborne outbreak of E. coli 015 7:H7 in
US. history. It took the lives of two fair attendees and sent 71 others to the hospital. An
investigation identified 781 persons with confirmed or suspected illness related to this

outbreak. The outbreak is thought to have been caused by contamination of the Fair’s

22

Well 6 by either a dormitory septic system or manure runoff from the nearby Youth
Cattle Barn GBopp 2003; Ackman 1997). More recently, in May 2000, an outbreak of E.
coli 015 7:H7 in Walkerton, Ontario resulted in at least seven deaths and 1,000 cases of
intestinal problems; public health officials theorize that one possible cause was
floodwaters washing manure contaminated with E. coli into the town’s drinking water
well; an investigation is currently underway (Brooke 2000). An outbreak of E. coli
015 7:H 7 was reported in Walkerton, Ontario, Canada from well water potentially
contaminated by manure runoff (Kluger 1998). Cow manure has specifically been
implicated as a causative factor in the high bacteria levels and ensuing swimming
restrictions on Tainter Lake, Wisconsin (Behm 1989). Among the many outbreaks
reports, studies have been published fi'om outbreaks in Scotland (Coia 1998; License
2001), Missouri (Swerdlow 1992) and Idaho (Vane Every 1995). This is only a small
sample of such reports. Investigations in Ontario on well water from a farm has shown a
relationship between the indicator in the well water and illness (Jackson 1998; Raina
1999). Finally, studies have been published on the presistance of E. coli in a farm
environment and the risks of contact and the relation with illness (Rahn 1997; Jones

1999; O’Brien 2001; Locking 2001; Ogden 2002; Strachan 2002).

Human Sewage

Population growth in coastal areas is increasing at a rate double that of population
growth worldwide. It is estimated that billions of gallons of treated and untreated
wastewater are discharged daily into the world's coastal waters. In developing nations,

90% of untreated sewage fiom urban areas is dumped into streams and oceans (Crossette

23

1996). In addition, runoff from heavy rains can worsen water quality. Increased bacterial,
viral, and toxin contamination may be associated with watershed pollution, loss of
wetlands (which naturally filter out pollutants), and overfishing (which decreases
predation). Heavy loadings of organic and inorganic nutrients change the ecological
balance, stimulating nuisance organisms (Burkholder 1997) and in some cases affecting

the virulence of indigenous species (Bates 1991).

Combined sewer systems are sewers that are designed to collect rainwater runoff,
domestic sewage, and industrial wastewater in the same pipe. Most of the time, combined
sewer systems transport all of their wastewater to a sewage treatment plant, where it is
treated and then discharged to a water body. During periods of heavy rainfall or
snowmelt, however, the wastewater volume in a combined sewer system can exceed the
capacity of the sewer system or treatment plant. For this reason, combined sewer systems
are designed to overflow occasionally and discharge excess wastewater directly to nearby
streams, rivers, or other water bodies. These overflows, called combined sewer
overflows (CSOS), contain not only storm water but also untreated human and industrial
waste, toxic materials, and debris. They are a major water pollution concern for the
approximately 772 cities in the U.S.that have combined sewer systems. CSOS may be
thought of as a type of “urban wet weather” discharge. This means that, like sanitary
sewer overflows (SSOS) and storm water discharges, they are discharges from a
municipality's wastewater conveyance infrastructure that are caused by precipitation
events such as rainfall or heavy snowmelt. Properly designed, operated, and maintained
sanitary sewer systems are meant to collect and transport all of the sewage that flows into

them to a publicly owned treatment works (POTW). However, occasional unintentional

24

discharges of raw sewage from municipal sanitary sewers occur in almost every system,
also known as SSOS. SSOs have a variety of causes, including but not limited to severe
weather, improper system operation and maintenance, and vandalism. EPA estimates that
there are at least 40,000 8803 each year (U SEPA 2002d). The untreated sewage from
these overflows can contaminate our waters, causing serious water quality problems. It
can also back-up into basements, causing property damage and threatening public health.
CSOS and wet weather SSOS contain a mixture of raw sewage, industrial wastewater and
storm water, and have resulted in beach closings, Shellfish bed closings, and aesthetic

problems.

Causal Connection of Illness and Sewggg

The possible transmission of infectious disease via primary contact with domestic
sewage has been the subject of many published epidemiological studies (including:
Stevenson 1953; Moore 1959; Cabelli 1982; Seyfield 1985; Fattel 1987; Ferley 1989;
Cheunk 1990; Balarajan 1991; Alexander 1992; Fleisher 1993; Corbett 1993, F leisher
1996; Butler 1997; Haile 1999; Prieto 2001; Bonadonna 2002; Dwight 2002). In
addition, CSOS have been identified as a source of bacteria including E. coli (Burm 1966;
Burm 1967). In a study of recreational water, it was reported that 34.5% of the
gastroenteritis infection were directly linked to domestic sewage in the swimming waters
and there was illness even though that water was within the EPA levels as acceptable

(Fleisher 1998).

25

Tools for E. coli Risk Assessment
ModelinLof the Watershjd

Global Information Systems (GIS) have been around since the 19608 but this was
mainly a research tool used by the federal government and large research universities.
However, this has changed and GIS is available to a much larger audience and has
become easier to use. This change was made possible by other technology trends such as
increased computer power and graphics in personal computers, improved and robust
databases, and increased computer savvy in our current society. Now, GIS is a tool not
only for scientists at NASA or a major research institution but rather, is also a tool used
by municipal planners on a local level and a skill learned by many at the undergraduate
level at many colleges across the country. In addition, the general public can access
environmental and watershed data in easy to use GIS formats through the EPA webpages.

GIS combines layers of information about a place to give the researcher a better
understanding of that place. What layers are combined depends on the researchers
purpose such as: finding the best location of a new store, analyzing environmental
damage, viewing similar crimes in a city to detect a pattern, or understanding land use in
a watershed. Many examples of watershed GIS data exist on the web such as: the Red
Cedar hosted by Michigan State University (MSU 2002), and the Kalamazoo River
hosted by Western Michigan University (WMU 2002). The USEPA hosts the “Surf your
Watershed” system (USEPA 2002) on which any who accesses this system can find GIS
information on any watershed in the United States. These watershed GIS databases, like

most, on based on USGS topological data (U SGS 2000) to define the area of the

26

watershed and micro-watersheds with it. In addition, theses GIS databases have layers
that include waterbodies, roads, cities, and in some cases land use coverage data.

There are many examples of GIS used as a tool in the analysis of watersheds.
Moreover, GIS has been used to model the activities and systems of a watershed. Non-
point pollution in a watershed has been a natural to model using GIS tools. BASIN S
designed by the EPA (U SEPA 2002b) STREAM (Spatial Tools for River basins,
Environment and Analysis of Management options) (Schepel 1998), and SIMPLE
(Spatially Integrated Models for Phosphorus Loading and Erosion) (Komecki 1999) are
examples of popular models that have designed for the evaluation of non-point pollution.
Agricultural land-use has been studied for non-point run-off. AGNPS (Agricultural Non-
Point Source) is a model designed by the USDA (Grunwald 2000, USDA 2002) for use in
determining the impact of agricultural activities on a watershed. AGNPS has been used
to determine the agricultural impact on coastal and estuarine ecosystems (Choi 1999) and
has been modified to integrate ARC/INFO databases in order to evaluate non point
source problem areas (Liao 1997). In addition, models have been designed to study
impact of the use of buffer strip on the water quality (Tim 1994). GIS tools have been
used to evaluate non-point pollution of surface waters with phosphorus and nitrogen
(Carpenter 1998, Robinson 1993). Additional GIS methods have been employed to help
understand sediment loading from agricultural land use (Rudra 1999).

Of course agricultural lands have not been the only areas studied. Urban systems
also have an impact on a watershed and GIS tools have been employed to study these

impacts. Storm-water management systems and their impacts have been modeled using

27

GIS (Sharnsi 1996). Estimations of mass loading from these types of systems evaluated
(Wong 1997, Adamus 1995).

Larger scale studies have used the entire watershed in order to determine problem
areas and what corrective actions could be implemented. The optimization the mix of
Best Management Practices (BMP) to reduce the loading on a waterbody has been the
goals of some research (Sample 2001, Wang 2000). GIS has been used to manage the
ecosystems of a watershed (Crawford 1998).

The impact of land use on a watershed is an obivious use of GIS tools but only
recently have studies on this subject have been published. Correlation between water
quality using conductivity as the measure and urban land use was identified (Wang
1997). Other assessments of water quality and land use have determined with out
surprise that land use does impact water quality (Wang 2001, Bhaduri 2000). There is
very little in the literature on using GIS in a watershed to determine risk fi'om microbial
pathogens based on land use. There are published results investigating septic systems as
potential pollutant but this study used nitrate as its measure (Stark 1999). In Ontario
geographic distribution of E. coli 015 7:H 7 infection and was compared to cattle
population (Michel 1999). The results indicate that cattle density had a positive and
significant association with the incidence of reported cases. GIS has been used to model

and predict pathogen loading from livestock (Fraser 1998).

The US. Environmental Protection Agency’s water programs and their
counterparts in states and pollution control agencies are increasingly emphasizing
watershed and water quality-based assessment and integrated analysis of point and

nonpoint sources. Better Assessment Science Integrating point and Non-point Sources

28

(BASINS) is a system developed to meet the needs of such agencies. It integrates a
geographic information system (GIS), national watershed and meteorologic data, and
state-of-the-art environmental assessment and modeling tools into one convenient
package. Originally released in September 1996, with a second release in 1998 BASINS,
addresses three objectives: (1) to facilitate examination of environmental information, (2)
to provide an integrated watershed and modeling fiamework, and (3) to support analysis
of point and non-point source management alternatives (U SEPA 2002b) BASINS
supports the development of total maximum daily loads (TMDLS), which require a
watershed-based approach that integrates both point and nonpoint sources. It can support
the analysis of a variety of pollutants at multiple scales, using tools that range from

simple to sophisticated.

The Bacterial Indicator Tool is a spreadsheet that estimates the bacteria
contribution from multiple sources. Output firm the tool is used as input to the water
quality model in BASINS. The tool estimates the monthly accumulation rate of fecal
coliform bacteria on four land uses (cropland, forested, built-up, and pastureland), as well
as the asymptotic limit for the accumulation should no washoif occur. The tool also
estimates the direct input of fecal coliform bacteria to streams from grazing agricultural

animals and failing septic systems (U SEPA 2002b).

E. coli Detection Method_s
The original mTEC Agar enumeration method (Dufour 1981) for E. coli was

introduced by EPA in 1986 (U SEPA, 1986b). A revised method was developed in 1998

29

by the EPA and has been designated as the modified mTEC method. Both the mTEC and
modified mTEC Agar methods use the membrane filter procedure. The two membrane
filter methods provide a direct count of E. coli in water based on the development of

colonies that grow on the surface of the membrane filter (Difco Labs, Detroit MI).

The PCR is a rapid and reliable tool for the molecular-based diagnosis of a variety
of infectious diseases (F redricks 1999). PCR analysis for screening drinking water and
environmental samples has been reported (Tsen 1998; Campbell 2001) and has been
utilized to identify E. coli in primary water Specimens (F ricker 1994; Juck 1996; Noble
2001), stool specimens (Paton 1993; Yavzori 1998; Sinton 1998), and outbreaks (Huerta
2000; Lodge 2002). In contrast, isolation of E. coli 015 7:H7 from water and other
environmental samples is laborious. Culture is problematic due to large numbers of other
flora that either overgrow or mimic the non-sorbitol-ferrnenting E. coli 015 7:H7 (de Boer
2000). PCR methods are preferred to the traditional culture assays because the E. coli
015 7:H7 grows poorly or not at all at 445°C (Doyle 1984). Recently, immunomagnetic
separation (IMS) has helped improve recovery by providing an antibody-based
concentration procedure that uses magnetic beads coated with antibody against E. coli
0157. Although many report the usefulness of IMS for testing artificially contaminated
samples, few reports have documented the use of IMS with naturally occurring,
epidemiologically linked specimens (Cubbon 1996; Wright 1994). Furthermore, there are
no reports documenting the use of IMS in support of a waterborne outbreak investigation.
Pulsed-field gel electrophoresis (PFGE) is useful for subtyping E. coli 015 7:H7 isolates
during outbreak investigations (Akers 1998; Ammon 1999; Bender 1997). PFGE is

reproducible and has sufficient discriminatory power to allow detection of minor genetic

30

variations among isolates (Willshaw 1997; Gouveia 1998). New technologies such as
biosensors have been studied for use with pathogenic bacteria (Stokes 2001; Leonard
2003; Lang MJ in process). Several reviews on detection methods have published

(including: Grif 1998; Baker 1999; Lejeune 2001; Theron 2002; Leonard 2003)

Risk Assess_ment Models

The following model (Figure 2-1) is for risk assessment, which was developed by

Figure 2-1. Risk Assessment Model (EPA 2000).

 

Hazard Assessment Dose Response Exposure
Assessment Assessment

 

 

 

 

 

 

 

 

 

 

  
 

 
     

   
  
     

 
  

Technical Dose
Response
Characterization

Technical
Exposure
Characterization

Technical
Hazard
Characterization

  
 
 

     
 

 

/ W \

Integrative RISK
Analysis CHARACI‘ARIZATION
SUMMARY

\ Risk Characterization Process /

 

 

 

 

 

 

 

31

the National Academy of Science and since adopted by the EPA and is the standard for
assessing risk in the environment (U SEPA 2000). The standard model for risk that is
used for a chemical contaminant or a carcinogenic material can not be used for a
microbial. Both chemical compounds and carcinogenic materials have a defined amount
of material that is plugged into the risk assessment model. This defined amount will only
change if one or more the following conditions exist: the contaminant of concern is
continuing to enter the environment, or natural attenuation is breaking down the
contaminant of concern, or the contaminant of concern is being removed by remediation.
Any of these condition the rate of increase or decrease of the contaminant of concern can
be calculated. Therefore, the amount of the contaminant of concern can be determined
fairly accurately. Based on these accurate calculations, models of mass transfer and
transport, as well as, the risk assessment are useable tools. However, microbial
substances such as E. coli are living material. Microbes live, multiply, or die based on
the environmental conditions and nutrients that are available. Microbes can be passed or
transmitted from person to person or from animals to people. In addition to the many
vectors from which a microbe can enter a person or host, once in the host a small amount
of microbes, which would not cause illness, can multiply into thousands or more if the
conditions are right and then cause infection. None of these factors that are unique to
microbes are considered in standard risk assessment models and such models will not

accurately predict the risk from a microbe.

Models have been introduced to determine the risk from E. coli in water (Gofti
1999; Gale 2001; Balbus 2002). Models for estimating the incidence from of swimming

related illness have been presented (Wymer 2002). Other models for estimating risk have

32

included long-term exposure and microbial densities (Haas 1996; Pinsky 2000). Still
other models have concentrated on the vector pathway (Strachan 2001). Risk assessment
models using complicated mathematical equations, Monte Carlo simulations, and
calculus have been devised to mathematically describe the unique factors of micobes
such as secondary infection and microbe survival (Eisenberg 1996; Chick 2001). For
such complicated models, software will need to be developed that is user fiiendly, in
order to have a wide acceptance of these models. Finally dose-response models have
attempted to determine infectious doses of pathogenic microbes (Holcomb 1999; Teunis
1999; Haas 2000). All of these dose-response models conclude that more data and

studies are needed and that statistical fit of the models to all data sets does not exist.

Epidemiological Evidence

Indigamr Bgcterig

A causal connection between fecal coliform and gastrointestinal sickness was first
identified in 1953. (Stevenson 1953). In the late 19708, in a landmark prospective cohort
study, reported a linear relationship between the incidence of gastroenteritis among
swimmers and marine bacterial counts (Cabelli 1982). Between 1973 and 1978,
participants were recruited at beaches from three US. locations-New York, Lake
Pontchartrain, Louisiana, and Boston, Massachusett8--and were contacted by telephone
days after going to the beach. Swimming status was self-selected, not randomly assigned,
and symptoms were self-reported. The mean proportion of swimmers with
gastrointestinal symptoms was 6.8% versus 4.6% in non-swimmers. When enterococcal

concentrations were above 1/ 100 ml, relative risk increased linearly, reaching 4.0 with

33

concentrations of 1,000/ 100 ml (p < 0.001). In addition, the frequency of gastrointestinal
symptoms was inversely related to the distance from known sources of municipal

wastewater (Cabelli 1982).

E. coli is an important cause of bacterial waterborne infection in untreated and
recreational water (Cabelli 1982). Infection can be life-threatening, especially in the
young and in the elderly. It can cause bloody diarrhea and, if not treated promptly, can
result in kidney failure and death. In particular, E. coli 015 7:H7 is emerging as the
second most important cause of bacterial waterborne disease after Shigella Species, which
is associated with human feces. E. coli 015 7:H7 was unknown until 1982, when it was
associated with a multistate outbreak of hemorrhagic colitis (Riley 1983; Tarr 1985;
Mead 1999).

There has a large body of literature involving the epidemiology of illness from the
use of recreational water (including: Schroeder 1968; Rosenbarg 1977; Favero 1985;
Seyfield 1985; Fattel 1987; Ferley 1989; Cheunk 1990; Balarajan 1991; Jones 1991;
Alexander 1992; Corbett 1993; Kay 1994; F leisher 1998; Haile 1999; Lopez-Pila 2000;
Prieto 2001; Dwight 2002). These studies have been conducted all over the globe
including Afiica, the Medetranian, South America, New Zealand, China, and Australia
(Cheung 1991; Vonschimding 1992; Corbett 1993; Harding 1993; Butler 1997; Mcbride
1998; Mourino-Perez 1999; Bonadonna 2002; Dionision 2002; Daby 2002). An excellent
review of this body of literature has been published which included 37 studies. This
study concluded that a causal dose-related relationship between gastrointestinal
symptoms and recreational water quality measured by bacterial indicator counts (Pruss

1998). Athletes, including tri-athletes, white-water canoeists, and surfers, have been

34

studied and these studies also Show a causal dose-related relationship between
gastrointestinal symptoms and recreational water quality measured by bacterial indicator
counts (Fewtrell 1992; Harrington 1993; Medema 1995; Medema 1997; van Asperen
1998).

The indicator bacteria to determine risk for recreational use in freshwater is E.
coli (EPA 1989, WHO 1998, EEC1994). There has been debate on the use and accuracy
of indicator bacteria level to determine the risk of the use of a waterbody (Haas 2001;
Griffin 2001; Mugglestone 2001; Leclerc 2001; Efstratiou 2001; Edberg 2000; F igeras
1997). There is evidence of interference of other serotypes that cause inaccuracies in
reported E. coli levels using the traditional EPA mandated plate culture methods
(McLellan 2001). Studies have investigated the original EPA risk analysis and
determined bias errors under-estimate the risk (Fleisher 1991; Fleisher 1990). However
another study reports contrary results and that the bias in this original studies have bias
that over-estimates the risk (Wymer 2002). Other measures such a beta glusicomase
levels (an enzyme that E. coli produces), RNA bacteriophages, amino acid sequences as
determined by PCR, and a matrix of indicators have been studied for use as replacements
to the current indicators (Lee 1997; Lopez-Pila 2000; Conboy 2001). The underlying
barrier to implementing these newer methods which have been shown to be more
accurate, is that the current EPA mandated method is very inexpensive. The cost of
equipment, materials and training that is needed for these new methods is too expensive
for cash strapped health department and will not implement unless it mandated by the

EPA.

35

E. coli 015 7:H 7

In contrast to the numerous studies, cases and reports of health effects from
exposure to indicator bacteria in recreational-use water, there are a limited number of
such findings in literature in regards to pathogenic E. coli 01 5 7:11 7. One of the reasons
for this that E. coli 015 7:H7 was first identified in 1982 (Riley 1983; Tarr PI 1985) and
food was typically identified as the transmission route for illness (Griffin 1991, Philips
1999, Coia 1998). In addition it is very difficult to identify in surface water and was first
identified in water in 1989 (McGowan 1989). Since E. coli 015 7:H7 may have infection
dose of as low as 10 cells (Mead 1999; Philips 1999) this may be impossible to detect.
That being said there are a few epidemiological studies that report E. coli 015 7:H7 as the
pathogen that caused illness in swimmers.

The first case of illness caused by E. coli 015 7:H7 was reported in 1991 and
occurred in a lake near Portland, Oregon. In a case control study 21 persons were
identified with E. coli 015 7:H7 infections by stool or serum samples and 7 of these
persons required hospitalization (Keene 1994, Van E88 1999). All of the affected persons
were children with the median age of six years old and all illnesses were not associated
with food or beverage but all of the affected persons reported they had been swimming.

It was concluded in the case control studies that swimming was strongly associated with
infection (p < 0.015) and the affected persons were more likely than controls to have
reported swallowing water while swimming (Keene 1994). The reported levels of the
indicator bacteria revealed substantial fecal contamination during the time period of the
study, in addition, E. coli 015 7:H7 was identified in the lake. Even though as many as

10,000 visitors a weekend would use the park in the lake was located, no studies or

36

calculations to determine prevalence, odds ratio, or relative risk were carried out during
the 7 week study period. A rough estimate of the incidence of illness based on seven
weeks of weekend visitors would be less than 0.03% (Keene 1994).

In 1993, six children were identified with illness caused by E. coli 015 7:H7 in a
1.5-mile radius in southwest London. One of the children later died from the illness. The
affected children had play in a paddling pool in which levels of E. coli were identified but
E. coli 015 7:H7 was not identified in these pools (Brewster 1994; Hildebrand 1996).

In 1993, 4 children, aged 1.5 to 3.5 years, all living in one town in the
Netherlands, were admitted to our hospital with the diagnosis hemolytic uremic
syndrome within one week. All 4 patients bathed in the same, shallow, recreational lake
within a period of 5 days. E. coli 015 7:H7 was demonstrated in the fecal samples of 2
index patients (Cransberg 1996).

In 1994, an E. coli 015 7:H7 outbreak was reported in Dutchess County, New
York. There were 12 confirmed cases of illness and all of the affected persons were
under the age of 14 years old. The case control study concluded that the illness was not
associated with food or beverage but with swimming and the affected children were more
likely than the controls to have ingested lake water (Ackman 1997). E. coli 015 7:H7 was
confirmed in affected children by bacterial subtyping. Samples taken from the lake both
before and after the outbreak were analyzed for E. coli and all results from these samples
were below 70cfu/100ml which is considered safe for full body in the water (Ackman
1997).

In 1995, a cluster 12 confirmed cases in children of E. coli 015 7:H7 induced

illness was reported in northern Illinois (JAMA 1996). All of the affected children had

37

been swimming in a lake in a state park. The case control study concluded that the illness
was not associated with food or beverage but with swimming (Mudgett 1998). In
addition the case control study determined that taking water into the mouth, swallowing
water, and time spent swimming all were risk factors for illness. Water samples that were
tested during the study showed elevated levels of E. coli but were not tested for E. coli
01 5 7:H 7. Water samples and sediment samples tested after the exposure did not reveal
any E. coli 0157:H7 in any of the nearly 100 samples (Mudgett 1998).

In 1996, in Georgia it was reported that a two-year girl who was attending a party
was diagnosed with E. coli 015 7:H7 that was confirmed by cultured analysis. Upon
firrther investigation 18 persons out of a total of 51 people who attended the party,
developed gastrointestinal illness and 10 of these met the case definition of E. coli
015 7:H7 infection as identified by elevated IgM or IgG levels (Friedman 1999).
Swimming in the pool significantly increased the risk of E. coli 015 7:H7 illness and no
other exposures were identified with illness. Although the pool had little to no chlorine
in it, no E. coli 015 7:H7 was identified in the pool.

In 1997, five cases of E. coli 015 7:H7 induced infection were identified in
Finland. The affected children were between the ages of 3 and 8 years old, all had swum
in a swallow, and had ingested lake water (Paunio 1999). An additional eight secondary
cases were reported and all of these cases were either family members or caretakers of
the affected children. The secondary cases were transmitted via person to person from
the affected children. All cases were confirmed for 015 7:H7 via the analysis of stool

samples. Samples of the lake water did not yield any E. coli 01 5 7:H 7.

38

In 1998, the CDC reported that 42 confirmed cases of E. coli 015 7:H 7 infection
nationwide and of that number, only 7 cases were connected to swimming pool in
Marietta Georgia (CDC 1999). Of the 7 swimming related illnesses unfortunately one of
the illnesses resulted in a death. In 1999, an outbreak of E. coli 015 7:H7 infection was
reported involving children at a day care in California. Seven cases were identified and
in a cohort stud the cause was determined as swallowing water in a freshwater lake
(Feldman 2002).

Also in 1999, it was reported that 36 patients developed E. coli 015 7:H7
infection in the State of Washington. Of the 36 patients, 28 were swimmers and 8
patients were infected via person-to-person transmission (Samadpour 2002). Although
E.coli levels were with in acceptable levels for swimming, 01 5 7.1-] 7 was identified in the
lake and is suggested to have come from a duck. This shows that the indicator organism,
E. coli, may have limitations in determining the real pathogenic risk of the water that is to
be used for swimming. The researchers used a PCR enriched method to detect the
pathogen. This method was used instead of the traditional culture assays because the E.
coli 015 7:H7 grow poorly or not at all at 445°C (Doyle 1984).

In all these reports or studies that have identified E. coli 015 7:H7 illness related
to swimming, only one identified E. coli 015 7:H7 in the water that was the exposure. In
addition, prevalence, risk, and incidence could not be calculated. Although E. coli
01 5 7:1-1 7 infection from swimming is very rare and the reports indicate that children are
at much higher risk that adults. This higher risk may be from weaker immune systems of
children as compared to adults but more likely it is due to the facts that children are more

likely to ingest water while swimming and that children are more likely than adults to

39

defecate while swimming. One study estimated that about 8 percent of US. outbreaks of
Escherichia coli 015 7:H 7 between the years 1982 and 1996 among children occurred as
a result of swimming (Griffin 1991). All of these studies suggest that the source of the E.
coli 015 7:H7 came from a person who was swimming and left behind the pathogen. The
use of scaled rubber swimming tnmks on children maybe have the greatest impact in

lowering the risk of E. coli 015 7:H7 infections when swimming.

Public Policy to Reduce Risk in Recreational Water

There has world-wide debate in regards to public policy and public safety for
recreational-use water (Figueras 1997; Bluemnthal 2000; Henrickson 2001; Efstratiou
2001). This review will focus on domestic (United States) public policy and leave the
discussion of EEC and WHO policy to the above-mentioned articles.

One of the major goals of the Clean Water Act of 1972 and its amendments
(CWA) is to ensure that US. waters are safe for fishing and swimming (33 U.S.C.A. §§
1251 to 1387). The use of water quality indicators that accurately reflect "safety" is
essential for this goal. The CWA mandates the use of indicator organisms to determine
safe levels and CWA section 303(d) requires the reporting of “impaired waters.” In
addition, section 303, establishes the water quality standards and Total Maximtun Daily
Load (TMDL) programs. The TMDL program is required by the rules that amended the

CWA issued in 1985 and then subsequently amended in 1992 (U SEPA 2002).

TMDL is a calculation of the maximum amount of a pollutant that a waterbody
can receive while still meeting water quality standards, and then determines an allocation

of that amount to the pollutant's sources. Water quality standards are set by States,

40

Territories, and Tribes and must meet or exceed the Federal standards set forth in the
CWA. These entities identify the uses for each waterbody, for example, drinking water
supply, contact recreation (swimming), and aquatic life support (fishing), and the
scientific criteria to support that use. A TMDL is the sum of the allowable loads of a
single pollutant from all contributing point and non-point sources. The calculation must
include a margin of safety to ensure that the waterbody can be used for the purposes the
State has designated. Additionally, the calculation must also account for seasonal

variation in water quality.

The indicator organism mandated for the monitoring of pathogens in fresh water
is E. coli. If the level of this organism is consistently elevated from the standard, then the
waterbody is declared an “impaired water” as defined in CWA 303 (d) and a TMDL is
required to be developed and implemented for that waterbody. As of 2001, there are
5,512 reported “impaired waters” due to pathogens (over 13% all impairments reported
and this is the second highest impairment behind sediment which is first at just below

14%) (USEPA 2001).

The Beach Environmental Assessment Closure and Health (BEACH) Act was
signed into law in October 2000 and amended the CWA (U SEPA 2002c). This
legislation directs the EPA Administrator to conduct studies to review health risks; to
develop indicators to detect the presence of pathogens; to offer guidance for state-to-state
application of the revised water quality criteria; to publish and revise regulations
requiring the monitoring of coastal waters; to provide technical assistance to states for

uniform assessment and monitoring procedures; and to establish a national coastal

41

recreation water pollution occurrence database and a listing of communities complying
with regulations published pursuant to this legislation. In addition, the Beach Act
authorizes $30 million annually for grants that would help states, local governments and
Indian tribes to monitor beach waters and notify the public when beach water exceeds the
established criteria. Moreover, some of these grants are designated for the development

of improved detection of pathogens in water, both freshwater and marine water.

It is mandated that consistent national health standards for beach water be
established by 2004. Currently, the EPA has determined that E. coli level is the best
indicator of risk to swimmers in fresh water. An E. coli concentration of less then 126
CPU/100ml, calculated as a geometric mean over 30 days, is considered to be safe for
firll body contact recreational use of water (U SEPA 1986). If a waterbody has a
designated use of full body contract recreational use and the E. coli levels are consistently
elevated from the above described standard, then the waterbody is declared a “impaired
water” as defined in CWA 303 (d) and a TMDL is required to be developed and
implemented for that waterbody. Protocol for developing pathogen TMDL EPA Office of

Water January 2001

Many county health departments in the State of Michigan routinely collect water
samples at beaches to determine if the water is safe for swimming. Samples are generally
taken one foot below the surface in water that is between three and six feet in depth. The
analysis is performed in a laboratory using standard methods. E. coli bacteria are counted

and judged against standards established by state rules. Results from the method are

42

available after approximately 28 hours; so the results from this analysis are reported the
following day.

County health departments take a minimum of three samples each time a beach
area is monitored. The My geometric mean calculated from these samples must be
below 300 E. coli per 100 milliliters for the water to be considered safe for swimming.
One or two of the samples may be above 300, but if the daily geometric mean is below
300, the beach is not in violation of the water quality standard. A minimum of five
sampling events (consisting of at least three samples per event) must be collected within a
30-day period for the results to be considered a reliable indication of water quality. After
30 days, a geometric mean is calculated for all the individual samples collected within
that time frame. This m geometric mean must be below 130 E. coli per 100 ml for
the water to be considered safe for swimming. Although the method yields results that
are very accurate, the results are not reported in less that 24 hrs. Officials close beaches
based on old information that may or may not be the actual condition of the water at the

time of closure and leave prior groups exposed.

EPA in both its 1992 National Water Quality Inventory and its Report to Congress
noted that pollution from wet weather discharges is cited by many states as the leading
cause of water quality impairment. Based on their reports and other assessments, the EPA
has concluded that wet weather discharges from both point and nonpoint discharges are
one of the largest threats remaining to water quality, aquatic life, and human health that

exist today. Areas of needed research and interest include but are not limited to:

43

. Development of technologies for preventing toxic substances and pollutants from
entering the downstream storm or combined sewer/drainage systems.

. Development of monitoring methodologies to measure the characteristics and
impacts of wet weather flows.

. Development of high-rate and high-efficiency WWTP treatment technologies.

While not subject to secondary treatment requirements, CSOS must nevertheless meet
water quality-based and technology based standards under NPDES permits to comply
with the Clean Water Act. Based upon U.S. EPA’S 1989 CSO strategy and 1994 National
CSO Policy, CSO communities are required to implement nine minimum control
technologies, and develop a long-term CSO control plan to meet water quality standards.
The nine minimum controls are generally met through management of the existing CSS,
while the long term controls will involve capital improvements such as the retention and

treatment, or sewer separation.

EPA'S CSO Control Policy, published April 19, 1994, is the national framework for
control of CSOS. The Policy provides guidance on how communities with combined
sewer systems can meet Clean Water Act goals in as flexible and cost-effective a manner
as possible. EPA'S Report to Congress on implementation of the C80 Control Policy
assesses the progress made by EPA, states, and municipalities in implementing and
enforcing the C80 Control Policy. "Wet weather discharges" refers collectively to point
source discharges that result from precipitation events, such as rainfall and snowmelt.

Wet weather discharges include storm water runoff, combined sewer overflows (C SOS),

and wet weather sanitary sewer overflows (SSOS). Storm water nuroff accumulates
pollutants such as oil and grease, chemicals, nutrients, metals, and bacteria as it travels
across land. CSOS and wet weather SSOs contain a mixture of raw sewage, industrial
wastewater and storm water, and have resulted in beach closings, shellfish bed closings,

and aesthetic problems.

Properly designed, operated, and maintained sanitary sewer systems are meant to
collect and transport all of the sewage that flows into them to a publicly owned treatment
works (POTW). However, occasional unintentional discharges of raw sewage from
municipal sanitary sewers occur in almost every system. These types of discharges are
called sanitary sewer overflows (SSOS). 8808 have a variety of causes, including but not
limited to severe weather, improper system operation and maintenance, and vandalism.
EPA estimates that there are at least 40,000 $808 each year. The untreated sewage from
these overflows can contaminate our waters, causing serious water quality problems. It
can also back-up into basements, causing property damage and threatening public health.

Table 2-3. The Nine Minimum Controls for Combined Sewer Overflows

1. Proper operation and regular maintenance programs for the sewer system and the

CSOS

Maximum use of the collection system for storage

Review and modification of pretreatment requirements to assure CSO impacts are

minimized

Maximization of flow to the publicly owned treatment works for treatment

Prohibition of CSOS during dry weather

Control of solid and floatable materials in CSOS

Pollution prevention

Public notification to ensure that the public receives adequate notification of CSO

occurrences and CSO impacts

9. Monitoring to effectively characterize CSO impacts and the efficacy of CSO
controls

we

”>199?

45

Since 1994, the USEPA has enforced these stringent new rules concerning wet
weather issues. Under the NPDES permit program, there are three program areas that
address each of the wet weather discharges described above. NPDES requirements from
runoff from concentratrated animal feeding operations (CAFOS) are described in an
earlier section of this paper. These programs share a range of cross-cutting issues and
affect a similar group of stakeholders. EPA believes that wet weather discharges should
be addressed in a coordinated and comprehensive fashion to reduce the threat to water
quality, reduce redundant pollution control costs, and provide State and local
governments with greater flexibility to solve wet weather discharge problems. To identify
and address cross-cutting issues and promote coordination, EPA established the Urban

Wet Weather Flows Federal Advisory Committee in 1995.

EPA is proposing to clarify and expand permit requirements for 19,000 municipal
sanitary sewer collection systems in order to reduce SSOS. The proposed SSO Rule will
help communities improve some of the Nation's most valuable infrastructure - our
wastewater collection systems - by requiring facilities to develop and implement new

capacity, management, operations, maintenance and public notification programs.

The nation's federal and state regulatory systems for protecting environmental
health have failed to keep pace with the rapid growth of factory farms. When Congress
passed the Clean Water Act (CWA) (33 U.S.C.A. Sections 1251 to 1387) 30 years ago, it
had the foresight to identify feedlots as an industrial source of pollution and to require

that feedlots be regulated as strictly as other industries. However, EPA has failed to

46

enforce these statutory requirements and the implementation of the regulations has been
pockmarked with loopholes. According to a 1995 General Accounting Office Report, in
1992 only 30 percent of the 6,600 farms that were large enough to be subject to federal
permit requirements actually obtained a permit under the Clean Water Act (US GAO
1995). To a greater or lesser extent, states have attempted to step into the void created by
an ineffective federal approach. Unfortunately, the states as a whole have also been
ineffective in regulating AF 08 (U SGAO 1995). Pennsylvania, Colorado and Alabama
have no permitting program, though programs are in the works. Illinois regulates only
livestock operations with animal waste lagoons but not those with underground manure
storage tanks, which are now the norm in Illinois. Some of these have leaked. California's
Central Valley issues permits only after an operation is caught polluting (N RDC 1999)
EPA'S proposed regulatory changes affect the existing NPDES provisions and the
existing ELG for “feedlots.” The NPDES provisions define and establish permit
requirements for CAFOs and the ELG establish the technology-based effluent discharge
standard that is applied to CAFOS. Both of these existing regulations were originally
promulgated in the 19708. The EPA proposed revision of the CAFO regulations would
affect operations that confine cattle and calves, milking cows, hogs and pigs, and poultry,
including broilers, egg laying chickens, and turkeys. Businesses that contract out the
raising or finishing phase of production might also be affected by the proposed co-
permitting requirements in the proposed CAFO regulations. Affected businesses may
include meat packing plants and poultry processing firms. The EPA has proposed two
alternatives on defining CAFOS that will be regulated. The two co-proposed alternatives

include the “two-tier structure” that would define as CAFOs all AF 08 with more than

47

 

500 AU and the “three-tier structure” that would define as CAFOs all AF 08 with more
than 1,000 AU and any operation with more than 300 AU, if they meet certain “risk-
based” conditions, as defined in the in the proposed rules (U SEPA 2002). EPA estimates
that both proposed alternative structures would regulate about 12,660 operations with
more than 1,000 AU, accounting for operations with more than a single animal type. The
two-tier structure would also regulate an additional 12,880 operations with between 500
and 1,000 AU, for a total of 25,540 operations. Under the three-tier structure, an
estimated 39,330 operations would be subject to the proposed regulations (10 percent of
all AF 08), estimated as the total number of animal confinement operations with more
than 300 AU (US EPA 2001). These have completed the comment period and the final

rules will be promulgated in the near future.

Conclusions

Ever since fecal contamination of water was determined a human health risk, there
has always been a great deal of concern regarding the level of coliform bacteria counts in
water. Many bodies of water throughout the world are considered to have counts above
acceptable levels. The sources of these E. coli are thought to be fecal contamination from
humans, domestic animals and wildlife, as well as runoff fi'om agricultural land,
inadequate septic systems or sewer overflow. Physical factors, such as seasonal
variability, rainfall, river flow, nutrient and the survival of bacteria in water, have an
impact on E. coli levels in a waterbody. Tools exist to assist in determining the risk to

swimmers of bacteria illness in a watershed. Although risk assessment models have been

48

studied, an accurate dose-response model for illness to swimmers of bacteria illness in a
watershed is not available.

There has a large body of literature involving the epidemiology of illness from
the use of recreational water. Elevated E. coli levels have been identified as the cause of
the illness and E. coli levels have been mandated as the indicator bacteria for the
determination of water safety for swimmers. The validity of E. coli as an indicator of risk
has been debated and E. coli may not be an accurate indicator of pathogenic risk.

In contrast to the numerous studies, cases and reports of health effects from
exposure to indicator bacteria in recreational-use water, there are a limited number of
such findings in literature in regards to pathogenic E. coli 01 5 7:H 7. In all these reports
or studies that have identified E. coli 015 7:H7 illness related to swimming, only one
identified E. coli 015 7:H7 in the water that was the exposure. Although E. coli 015 7:H7
infection from swimming is very rare and the reports indicate that children are at much
higher risk that adults. This higher risk may be from weaker immune systems of children
as compared to adults but more likely it is due to the facts that children are more likely to
ingest water while swimming and that children are more likely than adults to defecate
while swimming. The use of sealed rubber swimming trunks on children maybe have
the greatest impact in lowering the risk of E. coli 015 7:H7 infections when swimming.

Public policy in both the United States and around the world has addressed the
risk to swimmers of illness from bacteria in water. Levels of indicator bacteria that are
acceptable as to the risk for swimmers have been established by USEPA, EEC, and
WHO. In addition to these policies, laws and guidelines have been established to

decrease the microbial input in a waterbody from sewage sources and agricultural

49

facilities. Many of these laws and guidelines are being reviewed and revised in order to

improve the public health of swimmers.

References

Ackers M L, Mahon BE, Leahy B, Goode B, Damrow T, Hayes PS, Bibb WF, Rice DH,
Barrett TJ, Hutwagner L, Griffin PM, Slutsker L. 1998. An outbreak of Escherichia coli
015 7H 7 infections associated with leaf lettuce consumption. J. Infect. Dis. 177:1588-
1593

Ackman DS, Marks, et al. 1997. Swimming-associated haemorrhagic colitis due to
Escherichia coli 015 7:H7 infection: Evidence of prolonged contamination of a fresh
water lake. Epidemiology and Infection 119(1): 1-8.

Adamus CL, Bergman MJ, 1995, Estimating non point source pollution loads with GIS
screening, Water Res. Bull. 31(4): 647-655.

Alexexander LM, Heaven A, Tennant A, Moore R. 1992. Symptomatology of children in
contact with sea water contaminanted with sewage. Journal of Epideiologic Community
Health 46:340-344.

Ammon AL, Petersen R, Karch H. 1999. A large outbreak of hemolytic uremic syndrome
caused by an unusual sorbitol-fermenting strain of Escherichia coli 015 7:H 7. J. Infect.
Dis. 179: 1274-1277.

Artz RRE, Killham K. 2002. Survival of Escherichia coli 0157 : H7 in private drinking
water wells: influences of protozoan grazing and elevated copper concentrations. F ems
Microbiology Letters 216(1) 117-122.

Atherholt TB, LeChevallier MW, Norton WD, Rosen J S. 1998. Effects of rainfall on
Giardia and Crypto. J. Amer. Water Works Assoc. 90(9): 66-80.

Baker KH, Herson D S. 1999. Detection and occurrence of indicator organisms and
pathogens. Water Environment Research 71 (5): 530-551.

Balarajan R, Raleigh VS, Yuen P. 1991. Health risks associated with bathing in sea
water. British Medical Journal 303: 1444-1445.

Balbus J M. Embrey MA. 2002. Risk factors for waterborne enteric infections. Current
Opinion in Gastroenterology 18(1): 46-50.

Barton K, Fuller D. 1995. Testing the waters V: Politics and pollution at US. beaches.
New York: Natural Resources Defense Council.

50

Bates SS, de Frietas ASW, Pocklington R, Quilliam MA, Smith JC, Worms J. 1991.
Controls on domoic acid factors influencing production by the diatom Nitzschia pungens
f. multiseries: nutrients and irradiance. J Fish Aquat Sci 48:1136-1144.

Bauwens L, De Meurichy W, Vercarnmen F. 2000. Isolation of Escherichia coli 0157
from 200 animals. Vlaams Diergeneeskundig Tijdschrift, 69(2): 76-79.

Behm, D. 1989. 111 waters: The fouling of Wisconsin’s lakes and streams. The Milwaukee
Journal Sentinel. Special report: a series of articles published November 5-10.

Bender J B, Hedberg CW, Besser JM, Boxrud DJ, MacDonald KL, Osterholm MT. 1997.
Surveillance by molecular subtype for Escherichia coli 015 7:H7 infections in Minnesota
by molecular subtyping. N. Engl. J. Med. 337:388-394.

Besser RE, Lett SM, Weber JT, Doyle MP, Barrett TJ, Wells JG, Griffin PM. 1993. An
outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli 01 5 7 :H 7 in
fresh-pressed apple cider. JAMA 269:2217-2220.

Beutin L, Geier D, Steinruck H, Zimmerman S, Gleier K. 1994. Virulence factors and
phenotypic traits of verotoxigenic strains of Escherichia coli isolated from human
patients in Germany. Medical Microbiology Immunology 183:13-21.

Beutin L, Bulte M, Weber A, Zimmerman S, Gleier K. 2000. Investigation of human
infections with verocytotoxin-producing strains of Escherichia coli (V TEC) belonging to
sergroup 0118 evidence for zoonotic transmission. Epidemiology and Infection 125(1):
47-54.

Bhaduri B, Harbor J, Engel B, Grove M. 2000. Assessing water-scale, long-terrm
hydrilogic impacts of land-use change using GIS_NPS model. Enviro. Manag, 26(6):
643-658.

Bielaszewska M, Schmidt H, Liesegang A, Prager R, Rabsch W, Tschape H, Cizek A,
Janda J, Blahova K, Karch H. 2000. Cattle can be a reservoir of sorbitol-fermenting Shiga
toxin-producing Escherichia coli 015 7:H-strain8 and a source of human disease. Journal
of Clinical Microbiology, 38(9): 3470-3473.

Blumenthal UJ, Mara D D, Peasey A, Ruiz-Palacios G, Stott R. 2000. Guidelines for the
microbiological quality of treated wastewater used in agriculture: recommendations for
revising WHO guidelines. Bulletin of the World Health Organization 78(9): 1104-1116.

Bonadonna L, Briancesso R, Coccia AM, Semproni M, Stewardson D. 2002. Occurrence
of potential bacterial pathogens in coastal areas of the Adriatic Sea. Environ. Monitor and
Assess. 77(1): 31-49.

Bopp DJ, Saunders BD, Waring AL, Ackelsberg J, Dumas N, Braun-Howland E,
Dziewulski D, Wallace BJ, Kelly M, Halse T, Musser KA, Smith PF, Morse DL,

51

Limberger RJ. 2003. Detection, isolation, and molecular subtyping of Escherichia coli
015 7.'H 7 and Campylobacterjejuni associated with a large waterborne outbreak. Journal
of Clinical Microbiology 41(1): 174-180.

Briski F, Dsipos L, Petrovic M. 2000. Distribution of faecal indicatoribacteria and
nutrients in the Krka River in the region of the Krka National Park. Periodicum
Biologorum 102(3):273-28l.

Brooke, James. 2000. Few left untouched after deadly E. coli flows through an Ontario
town’s water. The New York Times. July 10.

Burkholder J M, Glasgow HB. 1997. Pfiesteria piscicida and other Pfiesteria-like
dinoflagellates: behavior, impacts, and environmental controls. Lirnnol Oceanogr (5 Part
2) 42:1052-1075.

Burm RJ, Vaughan RD. 1966, Bacteriological comparison between combined and
separate sewer discharges in southeastern Michigan. Journal of Water pollution Control
Federation 38(3): 400-409.

Burm RJ. 1967. The bacteriological effect of combined sewer overflows on the Detroit
River. Journal of Water pollution Control Federation 39(3): 410-424.

Bumens AP, Frey A, Lior H, Nicolet J. 1995. Prevalence and clinical significance of
Vero-toxin-producing Escherichia coli (V TEC) isolated from cattle in herds with and
without calf diarrhoea. J. Vet. Med. 42:311-318.

Butler T, F erson MJ. 1997. Faecal pollution of ocean swimming pools and storrnwater
outlets in eastern Sydney. Australian and New Zealand Journal of Public Health 21(6):
567-571.

Cabelli VJ, Dufour AP, McCabe LJ, Levin MA. 1982. Swimming-associated
gastroenteritis and water quality. Am J Epidemiol 115:606-616.

Cabelli VJ. 1989. Swimming-associated illness and recreational water-quality criteria.
Water Science and Technology 21(2): 13-21.

Campbell GR, Presser J, Glover A, Killham K. 2001. Detection of Escherichia coli

015 7:H7 in soil and water using multiplex PCR. Journal of Applied Microbiology 91(6):
1004-1010.

Carpenter SR, Caraco NF, Correll DL, Howarth RW, Sharply AN, Smith VH. 1998. Non-
point pollution of surface waters with phosphorus and nitrogen. Ecological Appl. 8(3):
559-568.

Carson RL. 1951. The Sea Around Us. New York: Oxford University Press.

52

Centers for Disease Control and Prevention. 1999. Letter to State and Territorial
Epidemiologists and Public Health Laboratory Directors, March 18, 1999.

CDC. 2001. Escherichia coli 01 5 7.'H 7.
htLp:/.r"www.cclc.gov/ncidod/dhmd/diseascinfo/escherichiacoli_g.htm

Chalmers RM, Aird H, Bolton FJ. 2000. Waterbome Escherichia coli 0157. Journal of
Applied Microbiology 88: 124S-132S.

Cheung WHS, Chang KCK, Hung RPS, Kleevens J WL. 1990. Health-effects of beach
water-pollution in Hong-Kong. Epidemiology and Infection 105(1): 139-162.

Cheung WHS, Chang KCK, Hung RPS. 1991. Variations in microbial indicator densities
in beach waters and health-related assessment of bathing water-quality. Epidemiology
and Infection 106(2): 329-344.

Cheung WHS, Hung RPS, Chang KCK, Kleevens J WL. 1991. Epidemiologic-study of
beach water-pollution and health-related bathing water standards in Hong-Kong. Water
Science and Technology 23(1-3): 243-252.

Chick SE, Koopman J S, Soorapanth S, Brown ME. 2000. Infection transmission system
models for microbial risk assessment. The Science of the Total Environment 274: 197-
207.

Chinyu Su MD, Brandt LJ. 1995. Escherichia coli 015 7:H7 infection in humans. Annals
of Internal Medicine 123(9):698-714.

Choi KS, Blood E. 1999. Modeling developed coastal watersheds with the agricultural
non-point model. J. AM. Water Res. Ass 35(2) 233-244.

Cimolai N, Basalyga S, Mah DG, Morrison BJ, Carter JE. 1994. A continuing assessment
of risk factors for the development of Escherichia coli 015 7:H7-associated hemolytic
uremic syndrome. Clinical Nephrol 42:85-89.

Clarke RC, Wilson JB, Read JC, Renvvick S, Rahn K, Johnson RP, Alves D, Karmali
MA, Lior H, McEwen SA, Spika J, Gyles CL. 1994. Verocytotoxin-producing
Escherichia coli (V TEC) in the food chain: preharvest and processing perspectives, p. 17-
24. In Karmali MA, and Goglio AG (ed.), Recent advances in verocytotoxin-producing
Escherichia coli infections. Elsevier Science B.V., Amsterdam, The Netherlands.

Coia, J E. 1998. Clinical, microbiological and epidemiological aspects of Escherichia coli
0157 infection. FEMS Immunology and Medical Microbiology 20(1): 1-9.

53

Coia JE, Sharp J CM, Campbell DM, Cunnow J, Ramsay CN. 1998. Environmental risk
factors for sporadic Escherichia coli 0157 infections in Scotland: Results of a descriptive
epidemiology study. Journal of Infection 36(3): 317-321.

Conboy MJ, Goss MJ. 2001. Identification of an assemblage of indicator organisms to
assess timing and source of bacterial contamination in groundwater. Water Air and Soil
Pollution 129(1-4): 101-118.

Corbett SJ, Rubin GL, Curry GK, Kleinbamn DG. 1993. The health-effects of swimming
at Sydney beaches. American Journal of Public Health 83(12): 1701-1706.

Cransberg K, vandenKerkhof Jhct, Banffer JRJ, Stijnen C, Wemars K, vandeKar Ncaj,
Nauta J, Wolff ED. 1996. Four cases of hemolytic uremic syndrome - Source
contaminated swimming water? Clinical Nephrology 46(1): 45-49.

Crawford IM, Tim US, Jain DK, Liao H. 1998. Managing ecosystems in a watershed
context: Progress made and emerging role of integrated spatial information technologies.
GIS 98 GIS Technologies and their Environmental Applications WIT Press.
Southampton, Boston.

Crossette B. 1996. Hope and pragmatism, for UN. cities conferences. New York Times,
3 June A3.

Cubbon MD, Coia JE, Hanson MF,Thomson-Carter FM. 1996. A comparison of
immunomagnetic separation, direct culture and polymerase chain reaction for the
detection of verocytotoxin-producing Escherichia coli 0157 in human faeces. J. Med.
Microbiol. 44 :2 1 9-222.

Curriero F C, Patz JA, Rose JB, Lele S. 2001. The Association between extreme
precipitation and waterborne disease outbreak in the United States, 1948-1994. Am. J. of
Pub. Health, 91(8): 1194-99.

Daby D, Turner J, Jago C. 2002. Microbial and nutrient pollution of coastal bathing
waters in Mauritius. Environmental International 27(7): 555-566.

Davies CM, Long JAH, Donald M, Ashbolt NJ. 1995. Survival of fecal microorganisms
in marine and freshwater sediments. Applied and Environmental Microbiology 61(5):
1888-1896.

de Boer E, Heuvelink AE. 2000. Methods for the detection and isolation of Shiga toxin-
producing Escherichia coli. J. Appl. Microbiol. 88:133S-143S.

Dionisio LPC, Garcia-Rosado E, Lopez-Cortes L, Castro D, Borrego JJ. 2002.

Microbiological and sanitary quality of recreational seawaters of southern Portugal.
Water Air and Soil Pollution, 138(1-4): 319-334.

54

Dorfman M. 2002. Testing the waters XII: A guide to water quality at vacation beaches.
NRDC, 21 1 pp.

Dorfinan M. 2001. Testing the waters XI: A guide to water quality at vacation beaches.
NRDC, 276pp.

Dorfinan M. 2000. Testing the waters X: A guide to water quality at vacation beaches.
NRDC, l66pp.

Dorfman M. 1999. Testing the waters IX: A guide to water quality at vacation beaches.
NRDC, 153pp.

Douglas AS, Kurien A. 1997. Seasonality and other epidemiological features of
haemolytic uraemic syndrome and E. coli 0157 isolates in Scotland, Scott Med J.
42(6):166-71.

Doyle MP and Schoeni JL. 1984. Survival and growth characteristics of E. coli associated
with hemorrhagic colitis. Applied Environmental Microbiology, 48(4): 855-856.

Drasar BS, Hill MJ. 1974. Human intestinal flora. p. 36-43. Academic Press, Ltd.,
London, United Kingdom.

Dwight RH, Semenza J C, Baker DB, Olson BH. 2002. Association of urban runoff with
coastal water quality in Orange County, California. Water Environment Research 74(1):
82-90.

Edwards DR, Coyne MS, Daniel TC, Vendrell PF, Murdoch JF, Moore PA. 1997.
Indicator bacteria concentrations of two northwest Arkansas streams in relation to flow
and season. Transaction of the ASAE, 49(1): 103-9.

Edwards DR, Coyne MS, Vendrell PF, Daniel TC, Moore PA, Murdoch JF. 1997. Fecal
coliform and Streptococcus concentrations in runoff from grazed pastures in northwest
Arkansas. Journal of the American Water Resources Association 33(2): 413-422.

Efstratiou MA. 2001. Managing coastal bathing water quality: The contribution of
microbiology and epidemiology. Marine Pollution Bulletin 42(6): 425-432.

Eisenberg JN, Seto EYW, Olivieri AW, Spear RC. 1996. Quantifying water pathogen
risk in an epidemiological framework. Risk Analysis 16(4): 549-563.

Enriquez C, Nwachuku N, Gerba CP. 2001. Direct exposure to animal enteric pathogens.
Reviews in Environmental Health 16(2): 117-131.

Epstein PR, Ford TE, Colwell RR. 1993. Marine ecosystems. Lancet 342:1216-2219.

55

Fattal B. Pelegolevsky E, Yoshpepurer Y, Shuval HI. 1986 The association between
morbidity among bathers and microbial quality of seawater. Water Science and
Technology 18(11): 59-69.

Fattal B, Pelegolevsky E, Agursky T, Shuval HI. 1987. The association between seawater
pollution as measured by bacterial indicators and morbidity among bathers at
mediterranean bathing beaches of Israel. Chemosphere 16(2-3): 565-570.

Favero MS. 1985. Microbiologic indicators of health risks associated with swimming.
American Journal of Public Health 75(9): 1051-1054.

F egan N, Desmarchelier P. 2002, Comparison between human and animal isolates of
Shiga toxin-producing Escherichia coli 0157 from Australia. Epidemiology and
Infection 128(3): 357-362.

Feldman KA, Mohle-Boetani JC, Ward J, Furst K, Abbott SL, Ferrero DV, Olsen A,
Werner S. 2002. A cluster of Escherichia coli 0157: Nonmotile infections associated
with recreational exposure to lake water. Public Heath Reports 117(4): 380-385.

Ferley JP, Zmirou D, Balducci F, Baleux B, Fera P, Larbaigt G, Jacq E, Moissonnier B,
Blineau A, Boudot J. 1989. Epidemiological significance of microbiological pollution
criteria for river recreational waters. International Journal of Epidemiology 18(1): 198-
205.

Fewtrell L. Godfree AF. Jones F, Kay D, Salmon RL, Wyer MD. 1992. Health-effects of
white-water canoeing. Lancet 339(8809); 1587-1589.

Figueras MJ, Polo F, Inza I, Guarro J. 1997. Past, present and future perspectives of the
EU bathing water directive. Marine Pollution Bulletin 34(3): 148-156.

F leisher J M. 1990. The effects of measurement error on previously reported
mathematical relationships between indicator organism density and swimming-associated
illness - a quantitative estimate of the resulting bias. International Journal of
Epidemiology 19(4): 1100-1106.

Fleisher J M. 1991. A reanalysis of data supporting United States federal bacteriological
water-quality criteria governing marine recreational waters. Research Journal of the
Water Pollution Control Federation 63(3): 259-265.

Fleisher J M, Jones F, Kay D, Stanwellsmith R, Wyer M, Morano R. 1993. Water and
non-water-related risk-factors for gastroenteritis among bathers exposed to sewage-
contarninated marine waters. International Journal of Epidemiology 22(4): 698-708.

Fleisher J M, Kay D, Salmon RL, Jones F, Wyer MD, Godfree AF. 1996. Marine waters
contaminated with domestic sewage: Nonenteric illnesses associated with bather
exposure in the United Kingdom. American Journal of Public Health 86(9): 1228-1234.

56

Fleisher J M, Kay D, Wyer MD, Godfree AF. 1998. Estimates of the severity of illnesses
associated with bathing in marine recreational waters contaminated with domestic
sewage. International Journal of Epidemiology 27(4): 722-726.

Ferguson CM, Coote BG, Ashbolt NJ, Stevenson IM. 1996. Relationships between
indicators, pathogen and water quality in an estuarine system. Water Research 30(9):
2045-2054.

Fraser RH, Barten PK, Pinney DAK. 1998. Predicting pathogen loading from livestock
using a Geographical Information System-based delivery model. J. Environ Qual 27(4):
935-945.

F redricks DN Relman DA. 1999. Application of polymerase chain reaction to the
diagnosis of infectious diseases. Clin. Infect. Dis. 29:475-486.

Fricker EJ, Fricker CR. 1994. Application of the polymerase chain reaction to the
identification of Escherichia coli and coliforrns in water. Lett. Appl. Microbiol. 19:44-46.

Friedman MS, Roels T, et al. 1999. Escherichia coli 015 7: H7 outbreak associated with
an improperly chlorinated swimming pool. Clinical Infectious Diseases 29(2): 298-303.

Gale P. 2001. Developments in microbiological risks assessment for drinking water.
Journal of Applied Microbiology 91(2): 191-205.

Geue L, Segura-Alvarez M, Conraths FJ, Kuczius T, Bockemuhl J, Karch H, Gallien P.
2002. A long-term study on the prevalence of Shiga toxin-producing Escherichia coli
(STEC) on four German cattle farms. Epidemiol. Infect. 129(1):l73-185.

Gofti L, Zmirou D, Murandi FS, Hartemann P, Potelon JL. 1999. Waterbome
microbiological risk assessment: a state of the art and perspectives. Revue
D’epidemiologie et de Sante Publique 47(1): 61-73.

Grif K, Karch H, Schneider C, Daschner FD, Beutin L, Cheasty T, Smith H, Rowe B,
Dierich MP, Allerberger F. 1998. Comparative study of five different techniques for
epidemiological typing of Escherichia coli 0157, Diagnostic Microbiology and
Infectious Disease 32(3): 165-176.

Griffin PM, Tauxe RV. 1991. The epidemiology of infections caused by Escherichia coli
015 7:H 7, other enterohemorrhagic Escherichia coli, and the Associated Hemolytic
Uremic Syndrome. Epidemiologic Reviews 13: 60-98.

Griffin PM, Ostroff SM, et a1. 1988. Illnesses associated with Escherichia coli 015 7:H7
infections - a broad clinical spectrum. Annals of Internal Medicine 109(9): 705-712.

57

Griffin DW, Lipp EK, McLaughlin MR, Rose J B. 2001. Marine recreation and public
health microbiology: Quest for the ideal indicator. Bioscience 51(10): 817-835.

Gouveia S, Proctor ME, Lee MS, Luchansky JB, Kaspar CW. 1998. Genomic
comparisons and Shiga toxin production among Escherichia coli 015 7:H7 isolates from
a day care center outbreak and Sporadic cases in southeastern Wisconsin. J. Clin.
Microbiol. 36:727-733.

Gunwald S, Norton LD. 2000. Calibration and validation of a non-point source pollution
model. Agr. Water Mana. 45(1): 17-39.

Haas CN. 1996. How to average microbial densities to characterize risk. Water Research
30(4): 1036-1038.

Haas CN, Thayyar-Madabusi A, Rose JB, Gerba CP. 2000. Development of a dose-
response relationship for Escherichia coli 0157 : H7. International Journal of Food
Microbiology 56(2-3) : 153-159.

Haile RW, Witte J S, Gold M, Cressey R, McGee C, Millikan RC, Glasser A, Harawa N,
Ervin C, Harmon P, Harper J, Derrnand J, Alarnillo J, Barrett K, Nides M, Wang G Y.
1999. The health effects of swimming in ocean water contaminated by storm drain
runoff. Epidemlology 10(4): 355-363.

Hancock DD, Besser TE, Kinsel ML, Tarr PI, Rice DH, Paros MG. 1994. The prevalence
of Escherichia coli 015 7:H7 in dairy and beef cattle in Washington State. Epidemiol.
Infect. 113:199-207.

Hancock DD, Besser TE, LeJeune J, Davic M, Rice D. 2001. The control of VTEC in the
animal reservoir. International Journal of Food Microbiology 66(1-2): 71-78.

Harding WR. 1993. Fecal-coliform densities and water-quality criteria in 3 coastal
recreational lakes in the SW Cape, South-Africa. Water SA 19(3): 235-246.

Harrington JF, Wilcox DN, Giles PS, Ashbolt NJ, Evans JC, Kirton HC. 1993. The health
of Sydney surfers - an epidemiologic study. Water Science and Technology 27(3-4): 175-
181.

Harvell CD, Kim L, Burkholder J M, Colwell RR, Epstein PR, Grime DJ, Hofrnann EE,
Lipp EK, Osterhaus AD, Overstreet RM. 1999. Emerging marine diseases-~climate links
and anthropogenic factors. Science 285:1505-1510.

Health Ecological and Economic Dimensions of Global Change Program. Marine

Ecosystems: Emerging Diseases as Indicators of Change. Boston, MAzCenter for Health
and the Global Environment, Harvard Medical School. 1998.

58

Henrickson SE, Wong T, Allen P, Ford T, Epstein PR. 2001. Marine swimming-related
illness: Implications for monitoring and environmental policy. Environmental Health
Perspectives 109(7): 645-650.

Heuvelink AE, van den Biggelaar F, de Boer E, Herbes RG, Melchers WJG, Huis In 'T
Veld J HJ , Monnens LAH. 1998. Isolation and characterization of verocytotoxin-
producing Escherichia coli 0157 strains from Dutch cattle and sheep. Journal of Clinical
Microbiology, 36(4): 878-882.

Heuvelink AE. 1999. The occurrence of Shiga toxin-producing Escherichia coli in
humans and animals. Tijdschrift Voor Diergeneeskunde, 124(22): 671 -678.

Hilbom ED, Mermin JH, Mshar PA, Hadler JL, Voetsch A, Wojtkunski C, Swartz M,
Mshar R, Lambert-Fair MA, Farrar JA, Glynn MK, Slutsker L. 1999. A multistate
outbreak of Escherichia coli 015 7:H7 infections associated with consumption of
mesclun lettuce. Arch. Intern. Med. 159:1758-1764.

Hildebrand JM, Maguire HC, Hollirnan RE, Kangesu E. 1996. An outbreak of E. coli
0157 infection linked to paddling pools. Commun Dis Rep CDR Rev 6(2):R33-36.

Hoar BR, Atwill ER, Elmi C, Farver TB. 2001. An examination of risk factors associated
with beef cattle shedding pathogens of potential zoonotic concern. Epidemiology and
infection 127(1): 147-155.

Holcomb DL, Smith MA, Ware Go, Hung YC, Brackett RE, Doyle MP. 1999.
Comparison of six dose-response models for use with food-bome pathogens. Risk
Analysis 19(6): 1091-1 100.

Huerta, M., I. Grotto, M. Gdalevich, D. Mirnouni, B. Gavrieli, M. Yavzori, D. Cohen,
and O. Shpilberg. 2000. A waterborne outbreak of gastroenteritis in the Golan Heights
due to enterotoxigenic Escherichia coli. Infection 28:267-271.

Jackson SG, Goodbrand RB, Johnson RP, Odorico VG, Alves D, Rahn K, Wilson J B,
Welch MK, Khakhia R. 1998. Escherichia coli 015 7:H7 diarrhoea associated with well
water and infected cattle on an Ontario farm. Epidemiologic Infection 120: 17-20.

JAMA. 1996. From the Centers of Disease Control and prevention lake associated
outbreak of E. coli 015 7:H7-Illinois 1995. JAMA 275(24):]872-1873.

Jones F, Kay D, Stanwellsmith R, Wyer M. 1991. Results of the lst pilot-scale controlled
cohort epidemiologic investigation into the possible health-effects of bathing in seawater
at Langland Bay, Swansea. Journal of the Institution of Water and Environmental
Management 5(1): 91-98.

59

Jones DL. 1999. Potential health risks associated with the persistence of Escherichia coli
0157 in agricultural environments. Soil Use and Management, 15(2): 76-83.

Joseph SW, Ingram DT, Kaper J B. 2002. The epidemiology, pathogenicity and
microbiology of foodbome Escherichia coli 0157 : H7. Reviews in Medical
Microbiology 13(2): 53-62.

Juck D, Ingram J, Prevost M, Coallier J, Greer C. 1996. Nested PCR protocol for the
rapid detection of Escherichia coli in potable water. Can. J. Microbiol. 42:862-866.

Karch H, Bielaszewska M, Bitzan M, Schmidt H. 1999. Epidemiology and diagnosis of
Shiga toxin-producing Escherichia coli infections. Diagnostic Microbiology and
Infectious Disease 34(3): 229-243.

Kariuki S, Gilks C, Kimari J, Obanda A, Muyodi J, Waiyaki P, Hart CA. 1999. Genotype
analysis of Escherichia coli strains isolated from children and chickens living in close
contact. Applied and Environmental Microbiology 65(2): 472-472.

Kay D, F leisher J M, Salmon RL, Jones F, Wyer MD, Godfree AF, Zelenauchjacquotte Z,
Shore R. 1994. Predicting likelihood of gastroenteritis from sea bathing - results from
randomized exposure. Lancet 344(8927): 905-909.

Keene WE, McAnulty J M, et a1. 1994. A swimming-associated outbreak of hemorrhagic
colitis caused by Escherichia coli 015 7:H7 and Shigella-Sonnei. New England Journal of
Medicine 33 1(9): 579-5 84.

Kerr M, Fitzgerald M, Sheridan JJ, McDowell DA, Blair IS. 1999. Survival of
Escherichia coli 0157 : H7 in bottled natural mineral water. Journal of Applied
Microbiology, 87(6) 833-841.

Kistemann T, Classen T, Koch C, Dangendorf F, F ischeder R, Gebel J, Vacata V, Exner
M. 2002. Microbial load of water reservior tributaries during extreme rainfall and runoff.
App and Environ, Microbio. 68(5):2188-97.

Kluger J. 1998. Anatomy of an outbreak, Time 152(5): 56-62.

Komecki TS, Sabbagh GJ, Storm DE. 1999. Evaluation of runoff, erosion, and
phosphorus modeling system- SIMPLE. J. AM. Water Res. A88. 35(4): 807-820.

Lacey RF, Pike EB. 1989. Water recreation and risk. Journal of the Institution of Water
and Environmental Management 3(1): 13-21.

Laegreid WW, Elder RO, Keen JE. 1999. Prevalence of Escherichia coli 0157 : H7 in
range beef calves at weaning. Epidemiology and Infection, 123(2) 291-298.

60

Lahti E, Keskimaki M, Rantala L, Hyvonen P, Siitonen A, Honkanen-Buzalski T. 2001.
Occurrence of Escherichia coli 0157 in Finnish cattle, Veterinary Microbiology, 79(3):
239-25 1 .

Law D, Kelly J. 1998. Use of heme and hemoglobin by Escherichia coli 0157 and other
Shiga-like-toxin-producing E. coli serogroups. Infect Irnmun 63:700-702.

Leclerc H, Mossel DAA, Edberg SC, Struijk CB. 2001. Advances in the bacteriology of
the Coliform Group: Their suitability as markers of microbial water safety. Annual
Review of Microbiology 55: 201-234.

LeJeune JT, Besser TE, Hancock DD. 2001. Cattle water troughs as reservoirs of
Escherichia coli 0157. Applied and Environmental Microbiology 67(7): 3053-3057.

Lee JV, Dawson SR, Ward S, Surman SB, Neal KR. 1997. Bacteriophages are a better
indicator of illness rates than bacteria amongst users of a white water course fed by a
lowland river. Water Science and Technology 35(11-12): 165-170.

LeJeune JT, Besser TE, Hancock DD. 2001. Cattle water troughs a8 reservoirs of
Escherichia coli 0157. Applied and Environmental Microbiology, 67(7) 3053-3057.

Leonard P, Hearty S, Brennan J, Dunne L, Quinn J, Chakraborty T, O'Kennedy R. 2003.
Advances in biosensors for detection of pathogens in food and water. Enzyme and
Microbial Technology 32(1): 3-13.

Liao HH, Tim US. 1997. An interactive modeling environment for non-point pollution
control. J. AM. Water Res. Ass. 33(3): 591-603.

Licence K, Oates KR, Synge BA, Reid TMS. 2001. An outbreak of Escherichia coli
015 7:H7 infection with evidence of spread from animals to man through contamination
of a private water supply. Epidemiology and Infection, 126(1) 135-138.

Lipp EK, Kruz R, Vincent R, Rodriguez-Palacios C, Farrah SR, Rose JB. 2001. The
effects of seasonal variability and weather on microbial fecal pollution and enteric
pathogens in a subtropical estuary. Estuaries 24(2): 266-76.

Locking ME, O’Brien SJ, Reilly WJ, Wright EM, Campbell DM, Coia JE, Browning
LM, Ramsay CN. 2001. Risk factors for sporadic cases of Escherichia coli 0157
infection: the importance of contact with animal excreta. Epidemiology and infection
127(2): 215-220.

Lodge FJ, Thompson DE, Call DR. 2002. PCR detection of specific pathogens in water: a
risk-based analysis. Environmental Science and Technology 36(12): 2754-2759.

61

Lopez EL, Diaz M, Grinstein S. 1989. Hemolytic uremic syndrome and diarrhea in
Argentine chilren: the role of Shiga toxins in children. Journal of Infectious Disease 160:

469-475.

Lopez-Pila JM, Szewzyk R. 2000. Estimating the infection risk in recreational waters
from the faecal indicator concentration and from the ratio between pathogens and
indicators. Water Research 34(17): 4195-4200.

MacDonald IAR, Gould IM, Cumow J. 1996. Epidemiology of infection due to
Escherichia coli 0157: a 3-year prospective study. Epidemiological Infection 116: 279-
284.

MacDonald JM, Ollinger ME, Nelson KE, Handy CR. 2000. Consolidation in US.
meatpacking. AER 785. Washington, DC: US. Department of Agriculture, Economic
Research Service. May 22.
http://wwwcrs.usda.gov/whatsncw/issucs/mcatpackimg,:’indcx.htm

Mallin MA, Ensign SH, McIver MR, Shank GC, Fowler PK. 2001. Demographic,
landscape, and meteorlogical factors controlling microbial pollution of coastal waters.
Hydrobiologia 460: 185-193.

Mariani-Kurkdjian P, Bingen E. 1999. Escherichia coli 0157 : H7, an emerging
pathogen. Presse Medicale, 28(37) 2067-2074.

Martin ML, Shipman LD, Wells JG, Potter ME, Hedberg K, Wachsmith IK, Tauxe RV,
Davis JP, Arnoldi J, Tilleli J. 1986. Isolation of Escherichia coli 015 7:H7 dairy cattle
associated with two cases of haemolytic uraemic syndrome. Lancet ii:1043.

Maule A, 2000, Survival of verocytotoxigenic Escherichia coli 0157 in soil, water and
on surfaces. Jouomal of Applied Microbiology 88:71 S-78S.

McBride GB, Salmond CE, Bandaranayake DR, Turner SJ, Lewis GD, Till DG. 1998.
Health effects of marine bathing in New Zealand. International Journal of Environmental
Health Research 8(3): 173-189.

McBride WD. 1997. Changes in US. livestock production, 1969-92. AER 754.
Washington, DC: US. Department of Agriculture, Economic Research Service. July.

McClure PJ, Hall S. 2000. Survival of Escherichia coli in foods. Journal of Applied
Microbiology 88 61S-70S.

McGee P, Bolton DJ, Sheridan JJ, Barley B, Kelly G, Leonard N. 2002. Survival of

Escherichia coli 0157 : H7 in farm water: its role as a vector in the transmission of the
organism within herds. Journal of Applied Microbiology, 93(4) 706-713.

62

McGowan KL, Wickersham E, Strockbin NA. 1989. Escherichia coli 015 7:H7 from
Water letter to editor. Lancet, 1(8644) 967-968.

McLellan SL, Daniels AD, Salmore AK. 2001. Clonal populations of thermotolerant
Enterobacteriaceae in recreational water and their potential interference with fecal
Escherichia coli counts. Applied and Environmental Microbiology 67(10): 4934-4938.

Mead PS, Griffm PM. 1998. Escherichia coli 0157 : H7. Lancet, 352(9135): 1207-1212.

Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee J S, Shapiro C, Griffin PM, Tauxe PV.
1999. F cod-related illness and death in the United States. Emerg. Infect. Dis. 5:1-32

Mechie SC, Chapman PA, Siddons CA. 1997. A Fifteen Month Study of E. coli 015 7:H7
in a Dairy Herd. Epidemiol. Infect. 118(1): 17-25.

Medema GJ, Vanasperen IA, Klokmanhouweling J M, Nooitgedagt A, Vandelaar MJ W,
Havelaar AH. 1995. The relationship between health-effects in triathletes and
microbiological quality of fresh-water. Water Science and Technology 31(5-6): 19-26.

Medema GJ, vanAsperen IA, Havelaar AH. 1997. Assessment of the exposure of
swimmers to microbiological contaminants in fresh waters. Water Science and
Technology 35(11-12): 157-163.

Michel P, Wilson JB, Martin SW, Clarke RC, McEwen SA. 1999. Temporal and
Geographical Distributions of Reported Cases of E. coli 015 7:H7 Infection in Ontario.
Epdemiol. Ifect. 122: 191-200.

Michino H, Araki K, Minarni S, Takaya S, Sakai N, Miyazaki M, Ono A, Yanagawa H.
1999. Massive outbreak of Escherichia coli 015 7:H7 infection in schoolchildren in Sakai
City, Japan, associated with consumption of white radish sprouts. Am. J. Epidemiol.
150:787-796

Minority Staff of the US. Senate Committee on Agriculture, Nutrition and Forestry,
Animal Waste Pollution in America: An Emerging National Problem, Washington, DC.
(December 1997), p3.

Miyagi K, Omura K, Ogawa A, Hanafirsa M, Nakano Y, Morimatsu S, Sano K. 2002.
Survival of Shiga toxin-producing Escherichia coli 015 7 in marine water and frequent
detection of the Shiga toxin gene in marine water samples from an estuary port.
Epidemiology and Infection 126(1) 129-133

Moake JL. 1994. Hemolytic Uremic Syndrome: Basis Science. Lancet 343:393-397.

Moore B. 1959. Sewage contamination of coastal bathing waters in England and Wales: a
bacteriological and epidemiological study. British Journal Hygiene 57:435—472.

63

Mourino-Perez RR. 1999. Studying the risks of ocean swimming. Epidemiology 10(4):
3 5 1-3 52.

Mudgett CC, Ruden R, et al. 1998. A beach-associated outbreak of Escherichia coli
015 7 .' H 7. Journal of Environmental Health 60(9): 7-13.

Mugglestone MA, Stutt ED, Rushton L. 2001. Setting microbiological water quality
standards for sea bathing - a critical evaluation. Water Science and Technology 43(12): 9-
18.

Nataro JP, Kaper JB. 1998. Diarrheagenic Escherichia coli. Clinical Microbiology
Reviews 11(1): 142-176.

National Institute of Health and Infectious Disease. 1996. Verotoxin-producing
Escherichia coli, January 1991-November 1995, Japan. Infectious Agents Surveillance
Report 17:1—2.

Niemi RM, Niemi J S. 1991. Bacterial pollution of waters in pristine and agricultural
lands. J. Environ. Qual. 20:620-627.

Nikiforuk A. 2000. When Water Kills. Maclean’s 113(24) 18-21.

Noble RT, F uhrrnan JA. 2001. Enteroviruses detected by reverse transcriptase
polymerase chain reaction from the coastal waters of Santa Monica Bay, California: low
correlation to bacterial indicator levels. Hydrobiologia 460: 175-184.

Noel J M, Boedeker EC. 1997. Enterohemorrhagic Escherichia coli: A family of
emerging pathogens. Digestive Diseases, 15(1-2) 67-91.

Notario R, Fain JC, Prado V, Rios M, Borda N, Gambande T. 2000. Animal reservoir and
genotipic characterization of Enterohemorrhagic Escherichia coli (EHEC) in Argentina,
Revista Medica De Chile 128(12): 1335-1341.

Obiri-Danso K, Jones K. 1999. Distribution and seasonality of microbial indicators and
thremophilic campylobacters in two freshwater bathing sites on the river lune in
northwest England. Journal of Applied Microbiology 87(6): 822-832.

O’Brien SJ, Adak GK, Gilharn C. 2001. Contact with farming environment as a major
risk factor for Shiga toxin (V cm cytotoxin)-producing Escherichia coli 015 7 infection in
humans. Emerging Infectious Diseases 7(6): 1049-1051.

Ogden ID, Hepburn NF, MacRae M, Strachan NJC, Fenlon DR, Rusbridge SM,
Pennington TH. 2002. Long-term survival of Escherishia coli 015 7 on pasture following
an outbreak associated with Sheep at a scout camp. Letters in Applied Microbiology
34(2): 100-104.

Olsen SJ, Miller G, Breuer T, Kennedy M, Higgins C, Walford J, McKee G, Fox K,
Bibb W, Mead P. 2002. A waterborne outbreak of Escherichia coli 015 7 : H7 infections
and hemolytic uremic syndrome: Implications for rural water systems. Emerging
Infectious Diseases 8(4): 370-375.

Osek J. 1999. Escherichia coli 015 7 - a dangerous pathogen with a broad spectrum of
illnesses. Medycyna Weterynaryjna, 55(4): 215-221.

Osek J, Gallien P. 2002. Molecular analysis of Escherichia coli 015 7 strains isolated
from cattle and pigs by the use of PCR and pulsed-field gel electrophoresis methods.
Veterinarni Medicina, 47(6): 149-158.

Padhye NV, Doyle MP. 1992. Escherichia-Cali 015 7-H7 - epidemiology, pathogenesis,
and methods for detection in food. Journal of Food Protection 55(7): 555-565.

Papadakis JA, Mavridou A, Richardson SC, Lampiri M, Marcelou U. 1997. Bather-
related microbial and yeast populations in sand and seawater. Water Research 3 1 (4): 799-
804.

Papanek PJ. 1994. How clean should the ocean be for swimmers. Lancet 343(8908):
1238-1239.

Paton AW, Paton J C, Goldwater PN, Manning PA. 1993. Direct detection of Escherichia
coli Shiga-like toxin genes in primary fecal cultures by polymerase chain reaction. J.
Clin. Microbiol. 31 :3063-3067.

Paunio M, Pebody R, Keskimaki M, Kokki M, Ruutu P, Oinonen S, Vuotari V, Siitonen
A, Lahti E, Leinikki P. 1999. Swimming associated outbreak of E. coli 01 5 7H 7.
Epidemiology of Infections, 122(1): 1-5.

Pavia AT, Nichols CR, Green DP, Tauxe RV, Mottice S, Greene KD, Wells JG, Siegler
RL, Brewer ED, Harmon D, Blake PA. 1990. Hemolytic-uremic syndrome during an
outbreak of Escherichia coli 015 7:H7 infections in institutions for mentally retarded
persons: clinical and epidemiologic observations. J. Pediatr. 116:544-551.

Pellizari VH, Pedroso DMM, Kirschner CC, Silva LAG, Martins MT. 1993. Assessment
of media using Beta-D-Glucuronidase activity for the detection of Escherichia-Cali in
Water. Revista De Microbiologia 24(3): 182-186.

Perez-Guzzi J I, Folabella A, Milwebsky E, Rivas M, F emandez-Pascua C, Gomez D,
Zamora A, Zotta C, Cordoba M. 2000. Isolation of Escherichia coli 015 7:H7 in storm
drainsin the sity of Mar del Plata with bacterial contarninantion of fecal origin. Rev
Argent Microbiology 32(3): 161-164.

Pettibone GW, Irvine KN. 1996. Levels and sources of indicator bacteria associated with
the Buffalo River area of concern. Journal of Great Lakes Research 22(4): 896-905.

65

 

Philips CA. 1999. The epidemiology, detection and control of Escherichia coli 015 7.
Journal of the Science of Food and Agriculture 79(11): 1367-1381.

Pickering LK, Obrig TG, Stapleton F B. 1994. Hemolytic-uremic syndrome and
enterohemorrhagic Escherichia- coli. Pediatric Infectious Disease Journal 13(6): 459-
475.

Pinsky PF. 2000. Assessment of risk from long-term exposure to waterborne pathogens.
Environmental and Ecological Statistics 7(2): 155-175.

Prieto MD, Lopez B, Juanes JA, Revilla JA, Llorca J, Delgado-Rodriguez M. 2001.
Recreation in coastal waters: health risks associated with bathing in sea water. quma] of
Epidemiology and Community Health 55(6): 442-447.

Pruss A. 1998. Review of epidemiological studies on health effects from exposure to
recreational water, Review of epidemic]. 27(1): 1-9.

Raina PS, Pcllari FL, Teare GF, Gcss MJ, Barry DAJ, Wilson JB. 1999. The relationship
between E-coli indicator bacteria in well- water and gastrointestinal illnesses in rural
families. Canadian qumal of Public Health-Revue Canadienne De Sante Publique 90(3):
172-175.

Rahn K, Renwick SA, Johnson RP, Wilson JB, Clarke RC, Alves D, McEwen S, Licr H,
Spika J. 1997. Persistence of Escherichia coli 015 7:H7 in dairy cattle and the dairy farm
environment. Epidemiological Infections 1 19:251-259.

Rasmussem MA, Casey TA. 2001. Environmental and food safety aspects of Escherichia
coli 015 7:H7 infections in cattle. Critical Reviews in Microbiology 27(2): 57-73.

Reilly WJ. 1997. E coli 015 7 in Scotland. Scottish Center for Infection and
Environmental Health Weekly Report.

Rice DH, Hancock DD, Besser IE. 2003. Faecal culture of wild animals for Escherichia
coli 015 7 : H7 152(3): 82-83.

Riley LW, Remis RS, Helgerson SD. 1983. Hemorrhagic colitis associated with a rare
Escherichia coli serotype. N Engl J Med;308:681-685

Rivas M, Balabi L, Miliwebsky ES, 1998. Sindrcme Urenricc Hemcliticc en Nincs de
Mendoza, Argentina. Medicina 58:1-7.

Robinson KJ, Ragan RM. 1993. Geographic Infcrrnaticn System based nonpoint
pollution modeling. Wat. Res. Bull. 29(6): 1003-1008.

66

Rose J B, Epstein PR, Lipp EK, Sherman BH, Bernard SM, Patz JA. 2001. Climate
variability and change in the Untied States: Potential impacts on water- and foodbcme
diseases caused by microbiologic agents. Environmental Health Perspectives 109: 211-
211S.

Rcsenburg ML, Koplan JP, Wachsmuth IK, Wells JG, Gangarcsa EJ, Guerrant RL, Sack
DA. 1977. Epidemic diarrhea at Crater Lake from entertcxigenic Escherichia coli: A
large waterborne outbreak. Annals of Internal Medicine 86: 714-718.

Rowe PC, Orrbine E, Licr H, Wells GA, Yetisir E, Clulcw M, McLaine PN. 1998. Risk
of hemolytic uremic syndrome after sporadic Escherichia coli 015 7 : H7 infection:
Results of a Canadian collaborative study. Journal of Pediatrics 132(5): 777-782.

Rudra RP, Dickinson WT, Abedini MJ, Wall GJ. 1999. A multi-tier approach for
agricultural watershed management. J. AM. Water Res. Ass. 35(5): 1159-1170.

Samadpour M, Stewart J, Steingart K, Addy C, Louderback J, McGinn M, Ellington J,
Newman T. 2002. Laboratory investigation of an E. coli 015 7:H7 outbreak associated
with swimming in Battle Ground Lake, Vancouver, Washington. Journal of
Environmental Health, 64(10): 16-20.

Samadpour M, Kubler M, Buck F C, Depavia GA, Mazengia E, Stewart J, Yang P, Alfi
D. 2002. Prevalence of Shiga toxin-producing Escherichia coli in ground beef and cattle
Feces from King County, Washington, Journal of Food Protection 65(8): 1322-1325.

Sample DJ, Heaney JP, Wright LT, Kcustas R. 2001. Geographic Information Systems,
decision support systems, and urban storm water management. J. Water Res. Plan. and
Manag. ASCE. 127(3): 155-161.

Sargeant JM, Gillespie JR, Oberst RD, Phebus RK, Hyatt DR, Bohra LK, Galland J C.
2000. Results of a longitudinal study of the prevalence of Escherichia coli 015 7:H7 on
cow-calf farms. American Journal of Veterinary Research 61(11): 1375-1379.

Schepel M. 1998. STREAM, spatil tools for river basin, environment, and analysis of
mangement options. GIS 1998, GIS Technologies and their Environmental Applications
WIT Press Southampton, Boston.

Schroeder SA, Caldwell JR, Vemcn TM, White PC, Granger SI, Bennett JV. 1968. A
waterborne outbreak of gastroenteritis in adults associated with Escherichia coli. Lancet
7545: 737-740.

Seyfried PL, Tobin RS, Brown NE, Ness PF. 1985. A Prospective-study of swimming-

related illness, 1. Swimming-associated health risk. American Journal of Public Health
75(9): 1068-1070.

67

Seyfried PL. 1989. Bacterial indicators to estimate the health-hazards associated with the
use of swimming pools. Toxicity Assessment 4(3): 317-327.

Sharnsi UM. 1996. Storm water management implementation through modeling and GIS.
J. Water Res. Plan. and Manag. ASCE. 122(2): 114-127.

Sinton LW, Finlay RK, Hannah DJ. 1998. Distinguishing human fi'om animal faecal
contamination in water: a review. New Zealand Journal of Marine and Freshwater
Research 32(2): 323-348.

Slutsker L, Ries AA, Greene KD, Wells JG, Hutwagner L, Griffin PM. 1997. Escherichia
coli 015 7:H7 diarrhea in the United States: Clinical and epidemiologic features. Annals
of Internal Medicine 126(7): 505-516.

Slutsker L, Ries AA. 1998. A nationwide case-control study of Escherichia coli 015 7 :
H7 infection in the United States. Journal of Infectious Diseases 177(4): 962-966.

Smayda TJ, Shimizu Y. 1993. Toxic Phytoplankton Blooms in the Sea. Londcanlsevier.

Srihari KA, Rao DV. 1998. Relative efficacy of 4-methyl umbellifery] beta-D
glucuronide (MUG) growth media for the detection of Escherichia coli in processed
foods. Journal of Food Science and Technology-Mysore 35(4): 314-319.

Stark SL, Nuckols JR, Rada J. 1999. Using GIS to investigate septic systems sites and
nitrate pollution. J. Environmental Health, 61(8): 15-6.

Stephenson GR and Street LV. 1978. Bacterial variations in streams from a southwest
Idaho rangeland watershed. qumal Environ Qual. 7(1):]50-157.

Stevenson AJ. 1953. Studies of bathing water quality and health. American qumal of
Public Health 43:529-538.

Stoke DL, Griffin GD, Tuan VD. 2001. Detection of E-coli using a microfluidics-based
antibody biochip detection system. Fresenius Journal of Analytical Chemistry 369(3-4):
295-301.

Strachan MJC, Fenlon DR, Ogden ID. 2001. Modelling the vector pathway and infection
of htunans in an environmental outbreak of Escherichia coli 015 7. FEMS Microbiology
Letters 203 (1): 69-73.

Strachan MJC, Dunn Gm, Ogden ID. 2002. Quantitative risk assessment of human
infection from Escherichia coli 015 7 associated with recreational use of an animal
pasture. International Journal of Food Microbiology 75(1-2): 39-51.

Swerdlow DL, Woodruff BA, Brady RC, Griffin PM, Tippen S, Donnell HD, Geldreich
E, Payne BJ, Meyer A, Wells JG, Greene KD, Bright M, Bean NH, Blake PA. 1992. A

68

waterborne outbreak in Missouri of Escherichia coli 015 7:H7 associated with bloody
diarrhea and death. Ann. Intern. Med. 117:812-819.

Synge BA. 2000. Veterinary significance of verocytotoxin-producing Escherichia coli
015 7. World Journal of Microbiology & Biotechnology, 16(8-9): 725-732.

Tamblyn SJ, deGrosbois RJ, Taylor D, Stratton J. 1999. An outbreak of Escherichia coli
015 7:H7 infection associated with unpasteurized non-commercial, custom-pressed apple
cider—Ontario, 1998. Can. Commun. Dis. Rep. 25:113-120.

Tarr P]. 1995. Escherichia-Cali 015 7-H7 - Clinical, diagnostic, and epidemiologic
aspects of human infection. Clinical Infectious Diseases 20(1): 1-10.

Tarr PI, Neill MA. 2001. Escherichia coli 015 7 : H 7. Gastroenterology Clinics of North
America 30(3): 735-747.

Tanaka Y, Yamaguchi N, Nasu M. 2000. Viability of Escherichia coli 015 7 : H7 in
natural river water determined by the use of flow cytometry. Journal of Applied
Microbiology 88(2):228-236.

Teunis PFM, Nagelkerke NJD, Haas CN. 1999. Dose response models for infectious
gastroenteritis. Risk Analysis 19(6): 1251-1260.

Theron J, Cloete TE. 2002. Emerging waterborne infections: Contributing factors, agents,
and detection tools. Critical Reviews in Microbiology 28(1): 1-26.

Tim US, Jolly R. 1994. Evaluationg agricultural nonpoint source pollution using
integrated Geographical Information Systems and hydrologic/water quality model. J.
Envir. Qual. 23(1): 25-35.

Thomas A, Cheasty T, Frost JA, Chart H Smith HR, Rowe B. 1996. Verocytotoxin-
producing Escherichia coli, particularly serogrcup 015 7, associated with human
infection in England and Wales: 1992-4. Epidemiol. Infect. 117:1-10.

Tiedemann AR, Higgens DA, Quigley TM, Sanderson HR, Bohn CC. 1988. Bacterial
water quality responses to four grazing strategies-comparison to Oregon standards. J.
Environ Qual. l7(3):492-498.

Tsen HY, J ian L2. 1998. Development and use of a multiplex PCR system for the rapid
screening of heat labile toxin 1, heat stable toxin 1] and Shiga-like toxin I and 1] genes of
Escherichia coli in water. J. Appl. Microbiol. 84:585-592.

USDA. 1997. Agricultural Census data.

USDA Agricultural Census data collected 1997 published 1999 and data collected 1974
published 1976.

69

USEPA Human Health Risk Assessment Guidelines (1986), and supplement (August
2000)

USDA. 2002. AGNPS website http://www.mrsars.usda.gov/morriS/products‘auimshtm;
http://www.sedlab.olemi88.edu/agnps.html.

USEPA. 2001. http://www.epa.gov/OWOW/tmdl/index.html

USEPA. 2002a. Surf your Watershed httpz/t’wwwepagov/surf

USEPA. 2002b. Better Assessment Science Integrating point and Non-point Sources
(BASIN S) http://www.epa.gov/watersciencc/ftp/basins/systcm/BAS INS3/bit.htm

USGS. 2002. Geographic Information Systems

http://www. usgsgov/research/giS/titlehtmI

van Asperen IA, Medema G, Borgdorff MW, Sprenger MJ W, Havelaar AH. 1998. Risk
of gastroenteritis among triathletes in relation to faecal pollution of fresh waters.
International Journal of Epidemiology 27(2): 309-315.

Van E88 E, Harding AK. 1999. A bacteriological study of natural freshwater swimming
areas in Western Oregon. Journal of Environmental Health 61(9): 14-19.

Van Every LR. 1995. Groundwater as a vehicle for disease transmission in southeastern
Idaho: a case study. Journal of Environmental Health 58: 16-19.

Vemozy-Rozand C, Montet MP, Ray-Guenict S. 2002. E-coli 015 7 : H7 in water: a
public health problem. Revue De Medecine Veterinaire 153(4): 23 5-242.

Verweyen HM, Karch H, Brandis M, Zirnmerhackl LB. 2000. Enterohemorrhagic
Escherichia coli infections: following transmission routes. Pediatric Nephrology 14(1):
73-83.

Vold L, Sandberg M, Jarp J, Wastescn Y. 2001. Occurrence and characterization of
Escherichia coli 015 7 isolated from cattle in Norway. Veterinary Research
Communications 25(1): 13-26.

Vonschimding YER, Kfir R, Cabelli V, Franklin L, Joubert G. 1992. Morbidity among
bathers exposed to polluted seawater - a prospective epidemiologic-study. South African
Medical Journal 81(11): 543-546.

Wallace J S, Cheasty T, Jones K. 1997. Isolation of vero cytotoxin-producing Escherichia
coli 015 7 from wild birds. Journal of Applied Microbiology 82:399-404.

70

Wang GD, Zhao T, Doyle MP. 1996. Fate of enterohemorrhagic Escherichia coli
015 7:H7 in bovine feces. Applied and Environmental Microbiology 62(7): 2567-2570.

Wang GD, Doyle MP. 1998. Survival of enterohemorrhagic Escherichia coli 015 7 : H7
in water. Journal of Food Protection 61(6) 662-667.

Wang XH, White-Hull C, Dyer S, Yang Y. 2000. GIS-ROUT: a river for watershed
planning. Envir. and Planning 27(2): 231-246.

Wang XH, Yin ZY. 1997. Using GIS to assess relationship between land-use and water
quality at a watershed level. Enviro. Intern. 23(1): 103-114.

Wang X. 2001. Integrating water quality management and land-use planning in a
watershed context. J. Enviro. Manag. 61(1): 25-36.

Watanabe H, Wada A, Inagaki Y, Itoh K, Tamura K. 1996. Outbreaks of
enterohaemorrhagic Escherichia coli 015 7:H7 infection by two different genotype
strains in Japan. 1996. Lancet 348:831-832.

Waters JR, Sharp J C, Dev VJ. 1994. Infection caused by Escherichia coli 015 7:H7 in
Alberta Canada and in Scotland: 3 five year review 1987-1991. Clinicla Infectiuos
Disease 19:834-843.

Weiskel PK, Howes BL, Heufelder GR. 1996. Coliform contamination of a coastal
embayment: sources and transport pathways. Environ Sci Technol 30: 1 872-1881 .

Wells JG, Davis BR, Wachsmuth IK, Riley LW, Remis RS, Sokolow R, Morris GK.
1983. Laboratory investigation of hemorrhagic colitis outbreaks associated with a rare
Escherichia coli serotype. J. Clin. Microbiol. 18:512-520.

Wells JG, Shipman LD, Greene KD, Sowers EG, Green JH, Cameron DN, Downes F P,
Martin ML, Griffin PM, Ostroff SM, Potter ME, Tauxe RV, Wachsmuth IK. 1991.
Isolation of Escherichia coli serotype 015 7:H7 and other Shiga-like-toxin-producing E.
coli from dairy cattle. J. Clin. Microbiol. 29:985-989.

Wheeler D, Alexander LM. 1992. Assessing the Health Risks of Sea Bathing, Journal of
the Institution of Water and Environmental Management 6(4): 459-467.

Whipp SC, Rasmussen MA, Cray WC. 1994. Animals as a source of Escherichia-Coli
pathogenic for human- beings. Journal of the American Veterinary Medical Association
204(8): 1168-1175.

Wickens B. 2000. Tragedy in Walkerton, Maclean’s 113(23) 24-26.

71

Willshaw GA, Smith HR, Cheasty T, Wall PG, Rowe B. 1997. Vero cytotoxin-producing
Escherichia coli 015 7 outbreaks in England and Wales, 1995: phenotypic methods and
genotypic subtyping. Emerg. Infect. Dis. 3:561-565.

WMU. 2002. Kalarnazoc River Watershed hosted by Western Michigan University
http://www.wmich.edu/watershedinfo/indexl .htrnl

Wright DJ, Chapman PA, Siddons CA. 1994. Immunomagnetic separation as a sensitive
method for isolating Escherichia coli 015 7 from food samples. Epidemiol. Infect.
113:31-39

Wong KM, Strecker EW, Strenstrom MK. 1997. GIS to estimate storm-water pollutant
mass loadings. Journal of Envir. Eng. ASCE 123(8): 737-745.

Wyer MD, Fleisher J M, Gough J, Kay D, Merrett H. 1995. An Investigation into
parametric relationships between enterovirus and fecal indicator organisms in the coastal
waters of England and Wales. Water Research 29(8): 1863-1868.

Wymer LJ, Dufour AP. 2002. A model for estimating the incidence of swimming-related
gastrointestinal illness as a function of water quality indicators. Environmetrics 13(5-6):
669-678.

Yavzori M, Porath N, Ochana O, Dagan R, Orni-Wasserlauf R, Cohen D. 1998.
Detection of enterotoxigenic Escherichia coli in stool specimens by polymerase chain
reaction. Diagn. Microbiol. Infect. Dis. 31:503-509.

Younts S, Alccilja EC, Osburn WN, Marquie S, Grooms DL. 2002, Differentiation of
Escherichia coli 015 7 : H7 from non-015 7 : H7 E-coli serotypes using a gas sensor-
based, computer- controlled detection system. Transactions of the ASAE 45(5): 1681-
1685.

Zhao T, Doyle MP, Zhao P, Blake P, Wu FM. 2001. Chlorine inactivation of Escherichia
coli 015 7 : H7 in water. Journal of Food Protection 64(10): 1607-1609.

Zhou ZJ, Nishikawa Y, Zhu P, Hong S, Hase A, Cheasty T, Smith HR, Zheng MG,
Haruki K. 2002. Isolation and characterization of Shiga toxin-producing Escherichia coli
015 7 : H7 from beef, pork and cattle fecal samples in Changchun, China. Journal of
Veterinary Medical Science 64(11): 1041-1044.

72

Chapter 3
The Effect of Rainfall, River Flow and other Meteorological Aspects on E. coli

Levels in the Red Cedar River

Introduction

Ever since fecal contarninaticn of water was determined a human health risk, there
has always been a great deal of concern regarding the level of coliform bacteria counts in
water. Many bodies of water throughout the world are considered to have counts above
acceptable levels. The sources of these coliforms are thought to be fecal contamination
from humans, domestic animals and wildlife, as well as runoff from agricultural land,
inadequate septic systems or sewer overflow (EPA 2000). The indicator organism for
fecal contamination in fresh water is Escherichia coli (E. coli) (EPA 2003, WHO 2003).
Physical characteristics of the watershed maybe risk factors that affect the concentration
levels of E. coli. This study will examine the effect of the seasons, the flow rate of the
river, rainfall, air temperature, water temperature, intensity of the sun, humidity, soil
moisture, ground temperature, and wind speed as factors related to concentration levels of
E. coli.

The seasonal variation of E. coli concentration levels have been studied and
reported in the literature. Over a twelve-year period (1984-1995) a study in Scotland
determined that seasonality was a factor with the highest amount of hospital admissions
for illness due to E. coli in July and August (Douglas 1997). It should be noted that this
study was not limited to recreational use of water and showed the seasonal pattern only
for patients under 15 years old. In contrast, a study in Florida showed that concentration

of fecal indicator organisms were the highest during the late fall and early winter months,

73

which corresponds to the wet, weather months in Florida (Lipp 2001). In addition the
study determined that the levels of fecal indicators were significantly associated with
rainfall, stream flow, and temperature. Moreover, a study of two stearns in Arkansas over
a three year period concluded that concentrations of indicator bacteria increased with
increasing flow rates and seasonal effects were observed on the indicators with the
highest levels occurring during the summer months (Edwards 1997). Furthermore, in a
study of a dairy herd E. coli 015 7:H7 was found in 4.3% of the herd and E. coli
concentration levels peaked during the May to July timeframe but was not found in the
herd from November to May (Mechie 1997). However, a study in northwest England
reported that bacteria indicators showed no seasonal variations over a two-year period
(Obiri—Dansc and Jones 1999).

Other studies support the association of the levels of fecal indicators in a water
body and rainfall events. In a study that examined 99 samples from three tributaries that
contributed to different drinking water reservoirs showed that E. coli levels along with
other bacteriological parameters increased considerably during extreme rainfall events
(Kistemann 2002). In a similar study on Delaware River, increased concentrations of
Giardia, Cryptosporidium and other microorganisms were associated with rainfall
(Atherholt 1998). In a retrospective study on waterborne disease outbreaks in the US.
for 1948 to 1994, analyzed 548 reported outbreaks as documented by the USEPA
database. The results showed that 51% of waterborne disease outbreaks were preceded
by precipitation events above the 90th percentile and 67% by events above the 80th
percentile (Curriero 2001). This study concluded that there is a statistically significant

association between high intensity rainfall events and waterborne disease outbreaks.

74

Other studies also conclude that E. coli levels increase with rainfall (Briski F 2000;
Ferguson 1996; Atherholt 1998; Pettibone and Irvine 1996; Niemi and Niemi 1991;
Mallin 2001; Noble 2003). Recent study by Rose (2001) suggests that an increase in
rainfall or snowmelt increases the impact of diseases caused by microbiologic agents.

An increase in the flow of a river has been also related to an elevation in the
concentration levels of E. coli in the river. Early studies in Idaho and Oregon found that
fecal E. coli concentrations were higher during period of high flow than during period of
lower flow (Stephenson and Street 1978; Tiedemann 1988). These finding were
confirmed in a study of two rivers in Arkansas that concluded that concentrations of
indicator bacteria increased with increasing flow rates (Edwards 1997).

The survival characteristics of a mixture of five E. coli 0157:H7 strains
innoculated at 103 CF U/ml in filtered and autoclaved municipal water, in reservoir water,
and in water from two recreational lakes were determined for a period of 91 days and
stored at three different temperatures of 8, 15, or 25 degrees C. Greatest survival was in
filtered autoclaved municipal water and least in lake water. Regardless of the water
source, survival was greatest at 8 degrees C and least at 25 degrees (Wang 1998).
According to the United States Food and Drug Administration (FDA) E. coli grows
between the temperatures of2.5°C and 45° C (USFDA 1992).

An objective of this study is to determine if there is a seasonal effect on the
concentration levels of E. coli in the Red Cedar River watershed. Another objective is to
determine if there is relationship between the flow rate of the river and concentration
levels of E. coli. In addition, an objective of this study is to determine the statistical

significance of concentration levels of E. coli as related to the watershed physical

75

characteristics of rainfall, air temperature, water temperature, intensity of the sun,
humidity, soil moisture, ground temperature, wind speed, duck population, as well as
river flow. Finally, an objective of this study is to use the statistical data to model risk
factors to so that water safety can be determined by the changes in physical
characteristics which maybe determinable in real time. It is hypothesized, based on the
results reported in the literature, that the model risk factors are directly related to rainfall

and associated factors.

Methods and Materials
D_at_a

The location of the study site is on the Michigan State University (MSU) campus
in East Lansing, Michigan. Water samples from the Red cedar River were collected
weekly during the 32 weeks from April to November over a three-year period. The water
samples were taken from the Farm Lane Bridge, which crosses the Red Cedar River in
the center of campus. These were collected by the MSU university physician’s office and
sent to the Michigan Department of Environmental Quality (MDEQ) for the
determination of E. coli concentration levels in water. The detection method for the
analysis of E. coli used by MDEQ was EPA Method 1103.1, which is a mTEC Agar
enumeration method introduced by EPA in 1986 and modified in 1998 (Dufour 1981;
USEPA, 1986b; USEPA 1998). The river flow measurements were taken from the
United States Geological Service (U SGS) gauging station that is located about 300 yards

down river from the Farm Lane bridge.

76

 

Figure 3-1. View of the Farm Lane Bride over the Red Cedar River from the USGS
Gauging Station on the Campus of Michigan State University.

The weather data were collected from the archives of the MSU horticulture
research weather station located on campus about two miles south of the Farm Land
Bridge. Data that were collected and used in this study included air temperature,
humidity, soil moisture at two inch depth, soil temperature at two inch depth, solar
radiation, wind speed, and precipitation. Finally, water temperature data and duck count
data were collected by Jo Latrnore, Ph.D. candidate at MSU as part of a separate project

and were only available for a portion of this study.

77

Temperatue probe

 

Hort Research
weather station Area that ducks
were counted

Figure 3-2. Locations of where measurements were taken for this study.

The Red Cedar River arises in Cedar Lake in the south-central Lower peninsula of
Michigan and flows about 45 miles (80 kilometers) to its confluence with the Grand
River in the city of Lansing. It has 12 tributaries and drains a total area of about 472
square miles (22,000 hectares). The river provides mid-Michigan residents with
numerous recreational opportunities that include angling, canoeing, kayaking,
photography and bird watching. The river also serves as a source of water for the
irrigation of crops throughout the watershed, as well as, drainage for agricultural land.

The weekly samples were taken from the 5,400-acre MSU campus with agricultural

78

operations two miles south of the river, traditional campus setting along the river, and the

urban landscape of the city of East Lansing less than a quarter mile north of the river.

 

., ‘ g
Farm Lane . -.;_f...r.£ 1'
Bridge ' ‘ ' ‘

   

Figure 3—3. The Red Cedar River Watershed with The Michigan State University campus
outlined in red.

Statistical Analysis
For seasonal effects, correlations were analyzed between the three years of data.

The first step in the statistical analysis included assessment of the correlations between
the E. coli levels and the physical and meteorological measurements. All the data was
collected at the same frequency as the water samples for the E. coli concentration

measurements, weekly over the 32 weeks from April to November over the three-year

79

period with the exception of the water temperature and duck counts. For all data the
number of samples (n) is equal to 96 except for water temperature where the number of
samples is equal to 33 and duck count where the ntunber of samples is 40. Because of
the lack of normality in the distribution of E. coli, precipitation and flow rate data,
Spearman’s correlation coefficient was used in all the correlations.

Once the physical and meteorological variables that significantly correlated with
E. coli were identified based on the correlation analysis, they were included in a logistic
regression with E. coli level exceeding a certain thresholds a8 a dependent variable. Two
different logistic regressions were run with different thresholds, one threshold value at
the E. coli level of 300cfu/100m], which is the highest acceptable full body contact level,
and another at the E. coli level of 1000cfu/100m1, which is the highest acceptable partial
body contact level. Statistical analysis was completed using SPSS version 11 software

package, as well as, SAS version 6.1.

Results and Discussion
Seasonal Effects

The plot in Figure 3-4 shows the weekly levels of E. coli with each year of data
overlaid so that the weeks of the each year line up. Note that the peak levels for each
year occur on different weeks. Statistical analysis on this data set shows that there is no

correlation either using parametric (Pearson) or non-parametric (Spearrnan)

80

 

8000
7000
6000
5000
4000
3000

E. coli concentration

2000
1000

 

u-mmtxms—MIONODFOOION
F‘— NNN

CD"-
N NOD

FFF

Week Number

 

—- 2000 - 2001 M-, 2002

 

 

 

 

Figure 3-4. E. coli Concentration from April to November in the Red Cedar River

correlation coefficients between the three years of data. Based on these analyses, it is
concluded that there is no observed seasonal effect for E. coli levels from the Farm Lane
Bridge over the three-year period from 2000 to 2002.

The plot in Figure 3-5 shows the weekly value for the flow rate at the time that
the sampling was done for the E. coli analysis. In contrast to the E. coli levels, the flow
rate correlation using Spearman’s Rho, between the three years is statistically significant
at least p < 0.05 or better. From this analysis, it can be concluded that there is an
observed seasonal variation in the flow rate of the Red Cedar River at the Farm Lane

Bridge.

81

 

900
800
700
600
500
400
300
200
1 00

0

Flow Rate

 

FMIDNQFC‘OIDNCDF
s-s-v-I—s-

co IO N
N N N N
Week Number

0)
N

‘—
(0

 

— 2000 ~— 2001 —— 2002

 

 

 

 

Figure 3-5. Flow Rate from April to November in the Red Cedar River

82

 

 

 

 

 

 

 

 

 

7.5
T I
7.0 - — spring ab
52:: summer '
— fall I"
E 0.5 -
8 b iii]

. b' __ a 3.:
E b r
3 6.0 'l i-Ll 7:“ [gig

5.5 - er 3'5; b
t r .
l .--.
5.0 - a" -r- T
2000 2001 2002

" means within the same year followed by the same letter are not significantly different (p<0.05)

Figure 3-6. Mean Values of the Log-transformed E. coli Concentrations in Spring,
Summer and Fall of the Studied Years.

 

. - spring
6'0 a ' m summer
- fall

55 d
' I

so -
' I

A. I

loom")

4.5 ‘ i i b
4.0 ‘

3.5 ‘

 

 

 

 

 

 

3.0 . He

 

* means within the same year followed by the same letter are not significantly different (p<0.05)

Figure 3-7. Mean Values of the Log-transformed Flow Rates in Spring, Summer and Fall
of the Studied Years.

83

There was no overall seasonal effect in the studied three years of the E. coli
concentration (p<0.1). However, the interaction between year and season was statistically
significant indicating that season effects were different in different years. Indeed, in 2000
average E. coli concentration was significantly higher in fall than in spring or summer
(p<0.1). In 2001, spring concentration was significantly greater than summer, while there
was no difference between fall and summer or fall and spring (p<0.05). In 2002, fall had
significantly lower concentration than spring or summer, the latter two being not
significantly different from each other (p<0.01). The observed inconsistency in seasonal
effects indicates that in the studied area occurrence of the increased E. coli concentrations

cannot be directly associated with any particular season.

Year-to-Year Observations

Shown in Figure 3-8 are the mean values along with error bars indicating the 95%
confidence level for E. coli and flow for each year. Note that as flow varies so do E. coli
levels. In the year 2002 the mean flow rate dropped by about 75% from the previous
years and the mean E. coli level for the year 2002 dropped over 40% in comparison to the

previous years.

84

 

1600

 

1200-

1000-

 

 

 

D E. coli

 

 

CI -200 , Flow

 

2000 2001 2002

Year

Figure 3-8. Yearly Mean and Standard Deviation of E. coli Levels and Flow

Shown in Figure 3-9 are the mean values for E. coli, flow (shown as bars) and
rainfall (overlaid as lines) for each year. The rainfall data were collected over the day of
E. coli sampling along with the two days previous to sampling and data was calculated
and reported as a cumulative total over a 24, 48, or 72 hour period. Not surprisingly
rainfall is associated with the E. coli level as does the flow rate. In the year 2002 the
mean cumulative rainfall for a 72-hour period dropped by about 65% from the year 2000
and the mean E. coli level for the year 2002 dropped about 38% in comparison to the year
2000. In the year 2002 the mean flow rate dropped by about 75% from the year 2000. It
makes common sense that the flow rate is lower for dry years and the year 2002 was a

drought year with no rain for an 11-week period from June through August.

85

 

1000 05

     

900 -. ~» 0.45
800 _ .- 0.4
700 — -» 0.35
600 — —» 0.3
500 «— _. 0.25
400 -_ .. 0.2
300 » «- 0.15
200 - -- 0.1
100 __ —» 0.05
o o

 

 

 

 

 

2001 2002

 

l- E. coli - Flow Rain 24h Rain 48h +Rain1211j
Figure 3-9. Yearly Mean of E. coli Levels, Flow, and Rainfall.

 

Michigan law (Michigan Public Health Code, PA 368 of 1978) based on EPA
guidelines (EPA 2003) mandates that the E. coli concentration of 3000fu/100ml is the
highest acceptable full body contact level, which is considered a person putting their head
under water and would include swimming and water skiing. The law also mandates that
E. coli concentration of 1000cfu/100ml is the highest acceptable partial body contact level
which is considered putting a body part other than ones head into the water and would
include fishing, canoeing, kayaking or boating. Table 3-1 shows the number weeks for

each that the E. coli concentration was exceeding the regulated levels.

86

Table 3—1. Number of Weeks that E. coli Concentration Over Full (1000 CFU/ml) or
Partial (300 CFU/ml) Body Contact.

Year Weeks over 300cfu/100ml Weeks over 1000cfu/100ml
2000 21 (65.6%) 5 (15.6%)
2001 17 (53.1%) 7 (21.9%)
2002 12 (37.5%) 5 (15.6%)
Total (n=96) 50 (52.1%) 17 (17.7%)

It is interesting to note that the nrunber of weeks that were over 300cfu/100m1
decreased from the year 2000 to 2002 in a similar trend to that of rainfall and flow. This
may indicates that the E. coli concentration at this level is influenced by rainfall and flow.
In contrast, the number of weeks that were over 1000cfu/100ml do not follow these trend
and indicating that there maybe other facts other than rainfall and flow rate that influence

occurrence of the very high concentrations of E. coli.

Effect of All Individual Factors

Rainfall and flow were not the only data that was collected to be compared to
E. coli concentration. Data that was collected and used in this study included air
temperature, humidity, soil moisture at two inch depth, soil temperature at two inch
depth, solar radiation, wind speed, precipitation, water temperature data and duck count.
All the data was collected at the same frequency as the water samples for the E. coli
concentration measurements, weekly over the 32 weeks from April to November over a
three-year period with the exception of the water temperature and duck counts. For all

data the number of samples (n) is equal to 96 except for water temperature where the

87

number of samples is equal to 33 and duck count where the number of samples is 40. For
the impact of each physical and meteorological measurement on the E. coli levels,
correlations using Spearman’s rho analysis were run keeping the E. coli level as the
independent variable and all others as dependant variables. The results are shown in

Table 3-2.

Table 3-2. Statistical Analysis of E. coli Concentration vs. River Flow, Rainfall, Air
Temperature, Humidity, Soil Moisture, Ground Temperature, Solar Radiation, Wind
Speed, Water Temperature, and Duck Count. (n = 96)

Statistical Formula Flow Rain 24 hrs Rain 48 hrs. Rain 72 hrs.
Spearman’s rho 0.368" 0.349" 0.510" 0.512"
Statistical Formula Temp hi Temp low Hum hi Hum lo
Spearrnan’s rho 0.079 0.222" 0.233* 0.330“
Statistical Formula Soil mst hi Soil mst low Grad tmp hi Grnd tmp lo
Spearman’s rho 0.361" 0.262“ 0.157 0.166
Statistical Formula Solar Windspd Water tmp'I Ducks”
Spearman’s rho -0.016 0.035 -0.129 0.010

* Statistically significant at p < 0.05; ** Statistically significant at p < 0.01.
a Water temperature had it = 33
b Duck count had 11 = 40

From the statistical analysis, river flow, rain, humidity and soil moisture were
Significantly correlated to E. coli concentrations. This group of factors are all related to
each other since when is rains humidity increases, soil moisture increases and river flow
increases. It is not surprising that all the variables fi'om this group are significantly
correlated with E. coli levels. Ducks have been shown to be a source of E. coli and have

been shown as a causation of illness to swimmers (Ackman 1997; Samadpour 2002).

F rorn this study there is no correlation between E. coli concentration and the number of

88

ducks in the river upstream fi'om the sampling point. In general, temperature, whether
air, soil or water, did not correlate to E. coli levels with the exception of the low
temperature of the day that the sample was collected. Solar radiation and wind speed
were not significantly correlated with E. coli concentrations.

Correlograms were calculated for each variable in each year. They indicated that
there was no significant spatial correlation between the variables. Hence, it is assumed
that the E. coli, 72-hour total rainfall (rain72), and air temperature measurements collected
at weekly intervals are independent. Analysis of the partial Speannan rank correlation
coefficients with rain72 controlled produced no significant correlations between E.coli
and any of the variables directly or indirectly related to amounts of cumulative 72 hr
precipitation, including flow, 24-hour total rainfall (rain24), 48-hour total rainfall
(rain48), humidity and soil moisture. The variables that had significant effect on the
E. coli when the amounts of cumulative 72-hour precipitation have been controlled for
were air and soil temperature. Because of this observation and also because of the
discussed above absence of the of the increase in E. coli in spring when the overall flow is
higher than in the other seasons, we may conclude that although flow is significantly
positively correlated with E. coli, the correlation is rather spurious since the amounts of
cumulative precipitation in past 48 or 72 hours is the driving force of both increase in
flow and in E. coli concentration. Higher correlation between rain72 with E. coli than
those between flow and E. coli can also be partially explained by the fact that high
intensive rains not only produce high flow but also increase the runoff from the

watershed carrying soil, manure and other potential sources of E. coli. High flow, on the

89

other hand, may result from a sequence of relatively mild rains that did not increase
amount of runoff sediment.

Once correlations were identified, the variables that were significantly correlated
were then analyzed using a logistic regression to determine the risk factors of each
variable. Since river flow, rain, humidity and soil moisture were significantly correlated
to E. coli concentration they were used as the factors in the logistic regression. Two
different logistic regressions were run with different thresholds, one threshold value at
the E. coli concentration of300cfu/100ml the highest acceptable full body contact level
and 1000cfu/100ml the highest acceptable partial body contact level.

The logistic regression model for describing the probability of E. colt>300 as a

function of rain72:

 

log 7r>300(Ram72) = —0.79 + 1.6Rain72
1' ”>300(Rai"72)

The observed probabilities of E. colr>300 and those predicted by the model are shown in
Figure 3-10. The rain72 data were divided in several classes such that there is a sufficient
(>5) number of events with E. coh>300 in each class. The observed probabilities were
calculated by dividing the number of occurrences of E. coli >300 observed in each rain72
class by the total number of E. coli observations in this class.

Positive slope of the logistic regression equation (p<0.001) indicates that the
probability of E. coli levels above 300 increases as rain72 increases. The lowest
probability is observed at zero rain72 and is equal to approximately 30%. The probability
of E. colr>300 reaches 95% for rain72 values exceed 2.4 cm. For every 1 cm increase in

rain72 values the odds of observing E. coli concentrations >300 increase on average em =

90

5 times. The 95% confidence interval for the logistic regression coefficient for rain72 is
equal to 0.71 to 2.47, hence we may conclude that the odds of E. coli exceeding 300 are
increasing at least twofold and at most twelve-fold for each 1 cm increase in rain72. This
result is supported by requirements in southern California to post health warnings on

beaches following all storms in which the rainfall is 1.25 cm or higher (Noble 2003).

 

 

1.0 ~ -———-
P
[:1 observed ”a?

+ filled
0.8 r

0.6 -

0.4 -

Probability of E.coli>300

0.2 ~

 

 

 

 

 

 

 

 

 

 

 

 

0.0 ' V I I I I I I I I
0 0-0.25 0250.50 0.50-0.75 0.75-1.50 1.50-2.25 2.25-3.00 3.00-3.75 >175

Cumulative 72 hr precipitation, cm

Figure 3-10. Probability of E. coli >300 cfm/ 1 00m] Resulting from Logistic Regression
Based on Cumulative 72 Hour Rainfall.

The only two variables that were statistically significant in the logistic regression

for E. colr>300 were rain72 (p<0.001) and templo (p<0.05):

10 M = —1.57 +1.79Rain7 + 0.082Temp,,,
1— ”>300 (Ramn) 2

91

Plot of the predicted probability of E. colr>300 as a function of rain72 and low
temperature of the day of sampling (templo) is shown in Figure 3-11. At zero rain72, the
probability of E. colr>300 is equal to ~11% if the templo is -6.5 °C, however, the
probability increases to 50% as the templo increases to 18 °C. At lower templo (<0°C) the

95% probability of E. coli concentration>300 is reached at approximately 1.3 cm rain72,
however at templo>15°C degrees the 95% probability of E. coli >300 occurs at 0.25 cm

rain72. The 95% confidence intervals for the logistic regression coefficients for rain72

and templo are equal to 0.80-2.77 and 0.01-0.15, respectively.

 

 

 

 

- o.o
- 0.2

8 - 0.4

<2 0.6

= 0.8 :1 0.3

8, 1:3 to

u.l

'6 0.6

E

g 04 6°

9 .

2 ~

0.

    

Low 5
daily “Mp9

twp, °C

Figure 3-11. Probability of E. coli >300 cfin/100ml Resulting from Logistic Regression
Based on Cumulative 72 Hour Rainfall and Low Air Temperature.

92

Since only 17 observations existed with E. colt>1000, the number of independent
variables that could be used in the regression equation had to be limited to 3 (Stokes
2000). Hence, the logistic regression for >1000 was conducted using several
combinations of three independent variables. The only variable that was statistically

significant in the logistic regression was rain72 (p<0.01):

 

104 ”>'°°° (Ram?) ]= —2.55 + 0.91Rain72
1" ”>iooo (Ramn)

Plots of the observed probabilities of E. coli >1000 calculated for grouped rain72
data are shown in Figure 3-12. The lowest probability is observed at zero rain72 and is
equal to approximately 7%. The highest probability predicted by the model for the
observed data was equal to 82% at the highest rain72 value of 4.5 cm. For every 1 cm
increase in rain72 values the odds of observing E. coli concentrations >1000 increase on

average e0“9| = 2.5 times with 95% confidence interval 1.6 to 3.9 times.

93

probability e.coli>1000

 

0.8
0.6 t

g I: observed

1;" + fitted

8.

l“ 0.4 -

g

'5

on

n

2

°' 0.2 -
0.0 I I I I I I I

 

 

 

 

 

T
0 00.25 0.25-0.50 0.50-0.75 0.75-1.50 1.50-2.25 >225
Cumulative 72 hr precipitation. cm

Figure 3-12. Probability of E. coli >1000 cfm/100ml Resulting from Logistic Regression
Based on Cumulative 72 Hour Rainfall.

Conclusions

Based on these analyses, it is concluded that there is no observed seasonal effect
for E. coli levels from the Farm Lane Bridge over the three-year period from 2000 to
2002. From the statistical analysis using Spearman’s rho, river flow, rain, humidity, low
temperature of the sampling day and soil moisture were significantly (p<0.05) correlated
to E. coli concentrations.

Other studies support the association of the levels of fecal indicators in a water
body and rainfall events and the results presented in this study support these findings. In
addition, previous studies have shown that increased river flow has been related to
increases in the concentration levels of E. coli in a river and the results presented in this

study also support these findings. From this study there is no correlation between E. coli

94

concentration and the number of ducks in the river upstream from the sampling point.
Solar radiation and wind speed did not correlate to E. coli concentrations.

Using governmental guidelines for maximum E. coli level for the safe use of
water for recreation, statistical models were designed and it was concluded that the odds
of E. coli exceeding 300 cfu/100m1 are increasing at least twofold and at most twelve-fold
for each 1 cm increase in 72-hour total rainfall. It was also concluded that if the low
temperature of the sampling day was greater than 15°C degrees the 95% probability of
E. coli exceeding 300 cfu/100ml occurs at 0.25 cm 72-hour total rainfall.

The results that have been presented indicate that programs to control run off
from riparian land during a rain event will have the greatest impact on lowering risk of

sickness to swimmers from fecal contamination.

References

Ackman, D., S. Marks, et al. (1997). “Swimming-associated haemorrhagic colitis due to
Escherichia coli 0157:H7 infection: Evidence of prolonged contamination of a flesh
water lake.” Epidemiology and Infection 119(1): 1-8.

Atherholt TB, LeChevallier MW, Norton WD, Rosen J S, 1998, Effects of Rainfall on
Giardia and Crypto, J. Amer. Water Works Assoc. 90(9): 66-80.

Briski F, Dsipos L, Petrovic M, 2000, Distribution of Faecal Indicator Bacteria and
Nutrients in the Krka River in the Region of the Krka National Park, Periodicum
Biologorum 102(3):273-281.

Curriero F C, Patz JA, Rose J B, Lele S, 2001, The Association between Extreme
Precipitation and Waterbome Disease Outbreak in the United States, 1948-1994, Am. J.
of Pub. Health, 91(8): 1194-99.

Douglas AS, Kurien A, 1997, Seasonality and Other Epidemiological Features of

Haemolytic Uraemic Syndrome and E. coli 0157 Isolates in Scotland, Scott Med .1.
42(6): 166-71 .

95

Edwards DR, Coyne MS, Daniel TC, Vendrell PF, Murdoch JF, Moore PA, 1997,
Indicator Bacteria Concentrations of Two Northwest Arkansas Streams in Relation to
Flow and Season, Transaction of the ASAE, 49(1): 103-9.

Ferguson CM, Coote Bg, Ashbolt Nj, Stevenson IM 1996 “Relationships between
Indicators, Pathogen and Water Quality in an Estuarine System,” Water Research 30(9):
2045-2054.

Kistemann T, Classen T, Koch C, Dangendorf F, F ischeder R, Gebel J, Vacata V, Exner
M, 2002, Microbial Load of Water Reservior Tributaries During Extreme Rainfall and
Runoff, App and Environ, Microbio. 68(5):2188-97.

Lipp EK, Kruz R, Vincent R, Rodriguez-Palacios C, Farrah SR, Rose JB, 2001, The
Effects of Seasonal Variability and Weather on Microbial Fecal Pollution and Enteric
Pathogens in a Subtropical Estuary, Estuaries 24(2): 266-76.

Mallin MA, Ensign SH, McIver MR, Shank GC, Fowler PK, 2001, “Demographic,
Landscape, and Meteorlogical Factors Controlling Microbial Pollution of Coastal
Waters,” Hydrobiologia 460: 1 85-193.

Mechie SC, Chapman PA, Siddons CA, 1997, A Fifteen Month Study of E. coli 0157:H7
in a Dairy Herd, Epidemiol. Infect., 118(1): 17-25.

Niemi RM, Niemi J S, 1991 “Bacterial Pollution of Waters in Pristine and Agricultural
Lands,” J. Environ. Qual. 20:620-627.

Noble RT, Weisberg SB, Leecaster MK, McGee CD, Dorsey J H, Vainik P, Orozco-
Borbon V, 2003, Storm effects on regional beach quality along southern California
shoreline, J. Water Health 01:1 23-31.

Obiri-Danso K, Jones K, 1999, Distribution and Seasonality of Microbial Indicators and
Thremophilic Carnpylobacters in two Freshwater Bathing Sites on the River Lune in
Northwest England,” J of Applied Microbiology 87(6): 822-832.

Pettibone GW, Irvine KN, 1996 “Levels and Sources of Indicator Bacteria Associated
wiht the Buffalo River Area of Concern,” J of Great Lakes Research 22(4): 896-905.

Rose J B, Epstein PR, Lipp EK, Sherman BH, Bernard SM, Patz JA, 2001, “Climate
Variability and Change in the Untied States: Potential Impacts on Water- and F oodbome
Diseases Caused by Microbiologic Agents,” Environmental Health Perspectives 109:

21 1-2 1 1 S.

Sarnadpcur M, Stewart J, Steingart K, Addy C, Louderback J, McGinn M, Ellington j,
Newman T, 2002, Laboratory Investigation of an E. coli 0157:h7 Outbreak Associated
with Swimming in Battle Ground lake, Vancouver, Washington, Journal of
Environmental Heaalth, 64(10): 16-20.

96

Stephenson GR and Street LV (1978) Bacterial Variations in Streams from 3 Southwest
Idaho Rangeland Watershed. J. Environ Qual. 7(1):]50-157.

Stokes et al., 2000

Tiedemann AR, Higgens DA, Quigley TM, Sanderson HR, Bohn CC (1988) Bacteria]
Water Quality Responses to four Grazing Strategies-Comparison to Oregon Standards. J.
Environ Qual. l7(3):492-498.

USEPA 2003, Bacteria] Water Quality Standards for Recreational Waters (Freshwater
and Marine Waters) EPA-823-R-03-008.

USF DA 1992 Center for Food Safety and Applied Nutririon, F oodbome Pathogenic
Microorganisms and Natural Toxins Handbook (“Bad Bug Book”).

Wang, G. D.; Doyle, M. P, 1998, Survival of enterohemorrhagic Escherichia coli 0157 :
H7 in water, Journal of Food Protection 61(6) 662-667.

World Health Organization (2003), Guidelines for Safe Recreational Water
Environments, Volume 1: Coastal and Fresh Water, Geneva, Switzerland.

97

Chapter 4

The Effect of Rainfall, Nutrient Levels, and Land-use on E. coli Levels in the Red

Cedar River

Introduction

Escherichia coli (E. coli) is the type species of the genus Escherichia, which
contains mostly motile rod-like gram-negative bacilli within the family
Enterobacteriaceae and the tribe Escherichia. E. coli bacteria live in the digestive systems
of humans and other warm-blooded animals. E. coli can be found in the fecal flora of a
wide variety of animals including cattle, sheep, goats, pigs, cats, dogs, chickens, and gulls
(Hancock 2001; Niemis and Niemis 1991).

There are a variety of sources that contribute bacteria and other pathogens to the
surface water. These sources include illicit waste connections to storm sewers or roadside
ditches, septic systems, combined and sanitary sewer overflows, storm (rain) runoff, wild
domestic animal waste, and agriculture runoff. Each of these sources is related to a type
of land-use and the type of land-use may have an impact on the amount of E. coli and
other nutrients that are discharged into nearby surface water.

Population growth in coastal areas is increasing at a rate double that of population
grth worldwide. It is estimated that billions of gallons of treated and untreated
wastewater are discharged daily into the world's coastal waters. In developing nations,
90% of untreated sewage from urban areas is dumped into streams and oceans (Crossette
1996). Coastal ecosystems are under increasing stress from a variety of human activities

that cause increased pollution, floral and fauna] changes (V itousek 1997; Epstein 1998).

98

E. coli sources have been linked to land use, mainly agricultural operations. Cow manure
has specifically been implicated as a causative factor in the high bacteria levels and
ensuing swimming restrictions on Tainter Lake, Wisconsin (Behm 1989). Among the
many outbreaks reported, studies have been published from outbreaks in Scotland (Coia
1998; License 2001), Missouri (Swerdlow 1992) and Idaho (Vane Every 1995), all of
which were associated with agricultural activities.

In contrast, studies have shown that residential development has had effect of
elevated E. coli levels (Frenzel and Couvillion 2002; Smith 2001; Mallin 2000). It has
been suggested that impervious surface may have be the cause for the elevated E. coli
levels in the urban areas (Mallin 2000). However, a study of an urban area to determine
sources of fecal contamination in river determined that the source was domesticated
animals and wildlife (Murry 2001). In a study of pristine lands in the fiords of
Scandinavia, it has been shown that fecal concentration can be above safe levels (N iemis
and Niemis 1991).

There are many examples of GIS used as a tool in the analysis of watersheds.
Moreover, GIS has been used to model the activities and systems of a watershed. Non-
point pollution in a watershed has been a natural to model using GIS tools. BASINS
designed by the EPA (U SEPA 2002b) STREAM (Spatial Tools for River basins,
Environment and Analysis of Management options) (Schepel 1998), and SIMPLE
(Spatially Integrated Models for Phosphorus Loading and Erosion) (Komecki 1999) are
examples of popular models that have designed for the evaluation of non-point pollution.
Agricultural land-use has been studied for non-point run-off. AGNPS (Agricultural Non-

Point Source) is a model designed by the USDA (Grunwald 2000, USDA 2002) for use in

99

determining the impact of agricultural activities on a watershed. AGNPS has been used
to determine the agricultural impact on coastal and estuarine ecosystems (Choi 1999) and
has been modified to integrate ARC/IN F O databases in order to evaluate non point
source problem areas (Liao 1997). In addition, models have been designed to study
impact of the use of buffer strip on the water quality (Tim 1994). GIS tools have been
used to evaluate non-point pollution of surface waters with phosphorus and nitrogen
(Carpenter 1998, Robinson 1993). Additional GIS methods have been employed to help
understand sediment loading from agricultural land use (Rudra 1999).

Of course agricultural lands have not been the only areas studied. Urban systems
also have an impact on a watershed and GIS tools have been employed to study these
impacts. Storm-water management systems and their impacts have been modeled using
GIS (Shamsi 1996). Estimations of mass loading fiom these types of systems evaluated
(Wong 1997, Adamus 1995). Larger scale studies have used the entire watershed in
order to determine problem areas and what corrective actions could be implemented. The
optimization the mix of Best Management Practices (BMP) to reduce the loading on a
waterbody has been the goals of some research (Sample 2001, Wang 2000). GIS has
been used to manage the ecosystems of a watershed (Crawford 1998).

The impact of land use on a watershed is an obvious use of GIS tools but only
recently studies on this subject have been published. Correlation between water quality
using conductivity as the measure and urban land use was identified (Wang 1997). Other
assessments of water quality and land use have determined with out surprise that land use
does impact water quality (Wang 2001, Bhaduri 2000). There is very little in the

literature on using GIS in a watershed to determine risk from microbial pathogens based

100

on land use. There are published results investigating septic systems as potential
pollutant but this study used nitrate as its measure (Stark 1999). In Ontario geographic
distribution of E. coli 0157:H7 infection and was compared to cattle population (Michel
1999). The results indicate that cattle density had a positive and significant association
with the incidence of reported cases. GIS has been used to model and predict pathogen
loading from livestock (Fraser 1998).

The relationship of nutrients levels to E. coli concentrations in surface water has
rarely been reported in the literature. Total suspend solid were strongly correlated to
indicator bacteria when flow rates were the highest (Pettibone and Irvine 1996). In
another study, a statistical analysis showed that there was a significant correlations
between levels of nitrate as well as levels of phosphate with the indicator bacteria fecal
coliform (Daby 2002). Heavy loadings of organic and inorganic nutrients can change the
ecological balance, stimulating nuisance organisms (Burkholder 1997) and in some cases
affecting the virulence of indigenous species (Bates 1991).

The objective of this study is determine if the type of land-use around a sampling
point is associated with elevated E. coli concentration levels. Another objective of this
study is to determine if the type of land-use around a sampling point is associated with
elevated levels of nutrients. In addition, an objective of this study is determine if there is
any association between the concentration levels of nutrients and the concentration levels
of E. coli. It is hypothesized that agricultural land-use has the highest discharges of E.
coli and nutrients and therefore has a greater negative impact to the Red Cedar River than

other types of land-use.

10]

Materials and Methods

The Red Cedar River arises in Cedar Lake in the south-central Lower peninsula of
Michigan and flows about 45 miles (80 kilometers) to its confluence with the Grand
River in the city of Lansing. It has 12 tributaries and drains a total area of about 472
square miles (22,000 hectares). The river provides mid-Michigan residents with
numerous recreational opportunities that include angling, canoeing, kayaking,
photography and bird watching. The river also serves as a source of water for the

irrigation of crops throughout the watershed, as well as, drainage for agricultural land.

A . ‘, '!
rid-‘3‘ a - " .3“, ._"
. . . j
"t ii 2 _ t.’ . .
I...” a.» . -. ,. “““
- r ; , > = e -.. 7.
i. ‘1 1 a 1 i. r ,, \-
g a“. A «:4 'f : .- _ I
~- a r I )..- . -
\cfii 5 1 , f u .
..- ',r' r- I; i v 6
. k. ‘ , . r”
:2; .,
\ ~ ‘4 , .‘ _ .
, 'A..- ‘f‘ ~
a“): .‘ ‘ t
; .
.— ' "Y.
i. g 1.1.1! I St
..A N i ..r’

Figure 4-1. The Red Cedar River Watershed with The Michigan State University
campus.

A data set that included 17 sampling points was studied over 13 sampling dates

over a 32-week period in 2001. The water samples for this multiple sampling point study

102

were collected by the Ingham County Health Department and the Livingston County
Health Department depending on the county in which the sampling point was located.

All water samples were sent to the MDEQ for the determination of E. coli levels using
EPA method 1103.] (Dufour 1981; USEPA, 1985; USEPA 1986; USEPA 2003). In this
multiple 8a outside lab were the standard methods that are recognized by the State of
Michigan for such water samples for analyzing these nutrients. In addition, rainfall data
from three different sites that include East Lansing, Williamston, and Howell so that the
rainfall data better represents each of the sampling sites. In addition, the 17 sampling
points were part of a larger 38 sampling point studies that collected E. coli data weekly
for 20 consecutive weeks (July to November) in 2000 and the 17 sampling point study,
the dataset includes results for the following nutrients: ammonia nitrate, total phosphorus,
and total suspended solids. The nutrient results were generated by an outside lab from
the water samples that were split with the MDEQ. The methods that were used by the for
32 consecutive weeks (April to November) in 2001.

The land use was determined by using the GIS program ArcView 3.1, (ESRI
Redland CA) with a National Land Cover Data (N LCD) layer for the watershed. Each
sampling point was located on the watershed map and a one-mile radius buffer was
created around the point. The land use was calculated on the semi circle of the buffer that
was upstream of the sampling point. If the land use for the calculated buffer was less
than 85% of one land use as categorized by the NLCD then the land use for that sampling
point was determined to be a mixed use. The land use variable for this analysis were
either urban, mixed, or agriculture. For the impact of each nutrient and rainfall

measurement, as well as, land use data on the E. coli levels, statistical analysis were

103

conducted using the Spearman’s rho keeping the E. coli levels as the independent
variable with rainfall, ammonia nitrate, total phosphorus, and total suspended solids set as
dependant variables. Statistical analysis was completed using SPSS version 11 software

package.

Results and Discussion

The graphical illustration shown in Figure 4-2 depicts the mean E. coli
concentration for the 17 sampling points. This represents 17 sampling points that were
studied over 13 sampling dates over a 32-week period in 2001 and has 221 E. coli
samples in the statistical analysis. From the analysis shown in Figure 4-2, mixed land use
has the highest impact on elevated E. coli levels and is followed by urban land use and
agricultural land use has the lowest impact on elevated E. coli levels. This difference is

Significantly different at p < 0.05.

104

2000.00-

1000.00-g

, 44;.» mm sjze

 

 

 

 

m7 ""5

 

 

 

 

 

 

 

0.00-f

 

l l l
agr mixed urb

Land Use

Figure 4—2. Means E. coli vs. Land use for 17 Selected Sampling Points
"‘ vertical bars represent plus or minus one standard error of the mean.

The graphical illustration shown in Figure 4-3 depicts the mean ammonia nitrate
concentration for the 17 sampling points. This represents 17 sampling points that were
studied over 13 sampling dates over a 32-week period in 2001 and has 221 ammonia
nitrate samples in the statistical analysis. From the analysis shown in Figure 4-3, mixed
land use has the highest impact on elevated ammonia nitrate levels and urban land use
and agricultural land use had similar low impact on elevated ammonia nitrate levels. This

difference is significantly different at p < 0.05.

105

20000-

 

t.ooot>- A
Amnon. ,
m 3 00071 cats-i 0100s

 

 

 

 

40000-

Mike

Figure 4-3. Mean Ammonia Nitrate vs. Land use for 17 Select Sampling Points

* vertical bars represent plus or minus one standard error of the mean.

The graphical illustration shown in Figure 4-4 depicts the mean total phosphorus
concentration for the 17 sampling points. This represents 17 sampling points that were
studied over 13 sampling dates over a 32-week period in 2001 and has 221 total
phosphorus samples in the statistical analysis. From the analysis shown in Figure 4-4,
mixed land use has the highest impact on elevated total phosphorus levels and is followed
by urban land use and agricultural land use has the lowest impact on elevated total

phosphorus levels. This difference is not significantly different.

106

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.1000-
0.0750-
0.0500-
Total H‘
Phos
_J_
00250-
0.0 70 0. 0.0440
m7 "'5
0.0000-
1 I I
nor M urb

Land Use

Figure 4-4. Mean Total Phosphorus vs. Land use for 17 Select Sampling Points
* vertical bars represent plus or minus one standard error of the mean.

The graphical illustration shown in Figure 4-5 depicts the mean total suspended
solids concentration for the 17 sampling points. This represents 17 sampling points that
were studied over 13 sampling dates over a 32-week period in 2001 and has 221 total
suspended solid samples in the statistical analysis. From the analysis shown in Figure 4-

5, agricultural land use has the highest impact on elevated total suspended solids levels

107

and is followed by urban land use and mixed land use has the lowest impact on elevated

total suspended solids levels. This difference is significantly different at p < 0.05.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12.0000— ——
t ___.___
80000- H
Total ; 1
Susp. ‘
Solid
.00...
9.6317 7.0900 7.7140
; or? n-5 n-5
00000- I T r
agr mixed urb
Land Use

Figure 4-5. Mean Total Suspended Solids vs. Land use for 17 Select Sampling Points
* vertical bars represent plus or minus one standard error of the mean.

The dataset included not only the results for ammonia nitrate, total phosphorus,
and total suspended solids but also rainfall. The rainfall data was collected over the day

of sampling along with the two days previous to sampling and data was calculated and

108

reported as a cumulative total over a 24, 48, or 72 hour period. This dataset represents 17
sampling points that were studied over 13 sampling dates over a 32-week period in 2001.
For the impact of each rainfall and nutrient measurement on the E. coli levels, statistical
analysis were conducted using the Spearman’s rho keeping the E. coli levels as the
independent variable with rainfall, ammonia nitrate, total phosphorus, and total
suspended solids set as dependant variables. The results are shown in Table 4-1.

Table 4-1. Statistical Analysis of E. coli Concentration vs. Rainfall, Ammonia Nitrate,
Total Phosphorus, and Total Suspended Solids (n = 221).

Statistical Formula Rain 24 hrs Rain 48 hrs. Rain 72 hrs.
Spearman’s rho 0.067 -0.012 0.084
Statistical Formula Am. Nitrate Total Phos. TSS
Spearman’s rho 0.274" 0.033 0.169*

* Statistically significant at p < 0.05; ““" Statistically significant at p < 0.01.

From the statistical analysis, ammonia nitrate was significantly correlated to E.
coli concentrations. The total suspended solids were significantly correlated to E. coli
concentrations. From this data, elevations in these nutrients do have a correlation with
elevated E. coli concentrations in the Red Cedar River.

It is interesting to note that rainfall did not Show any correlation using either
statistical analysis. This is contrary to other reported studies including a study during the
same time period on the Red Cedar River (Lang 2003). Other studies support the
association of the levels of fecal indicators in a water body and rainfall events. In a study
that examined 99 samples from three tributaries that contributed to different drinking

water reservoirs showed that E. coli levels along with other bacteriological parameters

109

increased considerably during extreme rainfall events (Kistemann 2002). In a related
study on Delaware River, increased concentrations of Giardia, Cryptosporidi um and
other microorganisms were associated with rainfall (Atherholt 1998). In a retrospective
study on waterborne disease outbreaks in the US. for 1948 to 1994, analyzed 548
reported outbreaks as documented by the USEPA database. The results from this study
showed that 51% of waterborne disease outbreaks were preceded by precipitation events
above the 90th percentile and 67% by events above the 80th percentile (Curriero 2001).
This study concluded that there is a statistically significance association between rainfall
and waterborne disease outbreaks. Other studies also conclude that E. coli levels increase
with rainfall (Briski 2000; Ferguson 1996; Atherholt 1998; Pettibone and Irvine 1996;
Niemi and Niemi 1991; Mallin 2001; Noble 2003). A recent study suggests that an
increase in rainfall or snowmelt increases the impact of diseases caused by microbiologic
agents (Rose 2001). In reviewing the data fi'om this study, the rainfall data was collected
from three different sites that include East Lansing, Williamston, and Howell so that the
rainfall data better represents each of the sampling sites. These rainfall collection sites
were not at the 17 sampling sites and probably did not represent the real rainfall that
occurred at the 17 sampling sites during the time period that the river water samples were
collected.

An expanded dataset that included the sampling sites to include all 38 sampling
Sites over the two-year period was evaluated to determine the impact of land use on E.
coli concentrations. The graphical illustration shown in Figure 4-6 depicts the mean E.
coli concentration for the 38 sampling points. This represents the 38 sampling points that

were collected E. coli data weekly for 20 consecutive weeks (July to November) in 2000

110

 

750.00%

 

500.00-é 4 ——

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Avg. __
E. coli
250.00 -,
473.07 454.21 387.00
R 081 0 [F14 7185
0.00 -’ I I I
agr mix urb

Land Use

Figure 4-6. Mean E. coli Levels vs. Land use for all 38 Sampling Points
"‘ vertical bars represent plus or minus one standard error of the mean.

and for 32 consecutive weeks (April to November) in 2001 and has 1976 E. coli samples
in the statistical analysis. From the analysis shown in Figure 4-6, agricultural land use
has the highest impact on elevated E. coli levels and is followed by mixed land use and
urban land use has the lowest impact on elevated E. coli levels but this is not significantly
different. This result is in contrast to the results shown in Figure 4-2 in which

agricultural land use had the lowest impact on elevated E. coli concentrations.

11]

Conclusions

Based on the results from this study, it is concluded that elevations of nutrients do
have an association with elevations in the levels of E. coli concentration. The
concentration levels of ammonia nitrate were significantly correlated to E. coli
concentrations and the levels of total suspended solids were significantly correlated to E.
coli concentrations.

These results indicate that if the locations of the sources of nutrients that enter
into nearby surface water are identified then proper best management practices (BMP)
can be implement at these locations and the implemented BMP will have an impact on
controlling nutrients into the surface water as well as having an effect to lower the E. coli
concentration in the surface water.

However, in this study rainfall does not show correlation with E. coli levels and
this is contradictory to what has been reported by others. This contradiction maybe
explained by the fact that the rainfall data was collected fi'om three different sites that
include East Lansing, Williamston, and Howell so that the rainfall data better represents
the sampling site. These rainfall collection sites were not at the 17 sampling Sites and
probably did not represent the real rainfall that occurred at the 17 sampling sites during
the time period that the river water samples were collected.

The impact of land use on E. coli concentrations was investigated but the results
did not reveal with any certainty as to which type of land-use has the greatest impact on
elevated E. coli concentrations. In the small sample that was analyzed, agricultural land
use had the greatest impact on elevated total suspended solids. However, mixed land use

had the greatest impact on elevated on ammonia nitrates. Larger sample sets need to be

112

studied to determine if the type of land-use does have an impact the concentration levels

of E. coli and other nutrients in nearby surface water.

References

Adamus CL, Bergman MJ. 1995. Estimating non-point source pollution loads with GIS
screening. Water Res. Bull. 31(4): 647-655.

Atherholt TB, LeChevallier MW, Norton WD, Rosen J S. 1998. Effects of rainfall on
Giardia and Crypto. J. Amer. Water Works Assoc. 90(9): 66-80.

Bates SS, de Frietas ASW, Pocklington R, Quilliam MA, Smith JC, Worms J. 1991.
Controls on domoic acid factors influencing production by the diatom Nitzschia pungens
f. multiseries: nutrients and irradiance. J Fish Aquat Sci 48:1136-1144.

Behm, Don. 1989. II] waters: The fouling of Wisconsin’s lakes and streams. The
Milwaukee Journal Sentinel. Special report: a series of articles published November 5-10.

Bhaduri B, Harbor J, Engel B, Grove M. 2000. Assessing water-scale, long-terrm
hydrilogic impacts of land-use change using GIS_NPS model. Enviro. Manag, 26(6):
643-658.

Briski F, Dsipos L, Petrovic M. 2000. Distribution of faecal indicator bacteria and
nutrients in the Krka River in the region of the Krka National Park. Periodicum
Biologorum 102(3):273-281 .

Burkholder J M, Glasgow HB. 1997. Pfiesteria piscicida and other Pfiesteria—like
dinoflagellates: behavior, impacts, and environmental controls. Limnol Oceanogr (5 Part
2) 42:1052-1075.

Carpenter SR, Caraco NF, Correll DL, Howarth RW, Sharply AN, Smith VH. 1998. Non-
point pollution of surface waters with phosphorus and nitrogen. Ecological Appl. 8(3):
559-568.

Choi KS, Blood E. 1999. Modeling developed coastal watersheds with the agricultural
non-point model. J. Am. Water Res. Ass. 35(2) 233-244.

Coia, J E. 1998, Clinical, microbiological and epidemiological aspects of Escherichia coli
0157 infection. FEMS Immunology and Medical Microbiology 20(1): 1-9.

Crawford IM, Tim US, Jain DK, Liao H. 1998. Managing ecosystems in a watershed
context: Progress made and emerging role of integrated spatial information technologies.

113

GIS 98 GIS Technologies and their Environmental Applications WIT Press
Southampton, Boston

Crossette B. 1996. Hope, and pragmatism, for UN. cities conferences. New York Times,
3 June2A3.

Curriero F C, Patz J A, Rose JB, Lele S. 2001. The association between extreme
precipitation and waterborne disease outbreak in the United States, 1948-1994. Am. J. of
Pub. Health, 91(8): 1194-99.

Daby D, Turner J, Jago C. 2002. Microbial and nutrient pollution of coastal bathing
waters in Mauritius. Environmental International 27(7): 555-566.

Dufour AP, Strickland ER, Cabelli VJ. 1981. Membrane filter method for enumerating
Escherichia coli. Appl Environ Microbiol 41:1152-1158.

Epstein PR, Ford TE, Colwell RR. 1993. Marine ecosystems. Lancet 342: 1216-2219.

Epstein PR. 1998. Marine ecosystems: emerging diseases as indicators of change.
Boston, MA:Center for Health and the Global Environment, Harvard Medical School.

Ferguson CM, Coote BG, Ashbolt NJ, Stevenson IM. 1996. Relationships between
indicators, pathogen and water quality in an estuarine system. Water Research 30(9):
2045-2054.

Fraser RH, Barten PK, Pinney DAK. 1998. Predicting pathogen loading fi'om livestock
using a geographical information system-based delivery model. J. Environ Qual 27(4):
935-945.

F renzel SA, Couvillion CS. 2002. F coal-indicator bacteria in streams along a gradient of
residential development. Journal of AWRA 38(1): 265-273.

Gunwald S, Norton LD. 2000. Calibration and validation of a non-point source pollution
model. Agr. Water Mana. 45( 1): 17-39.

Hancock, D. D., T. E. Besser, M. L. Kinsel, P. I. Tarr, D. H. Rice, and M. G. Paros. 1994.
The prevalence of Escherichia coli 0157:H7 in dairy and beef cattle in Washington State.
Epidemiol. Infect. 1 13:199-207.

Hancock DD, Besser TE, LeJeune J, Davie M, Rice D. 2001. The control of VTEC in the
animal reservoir. International Journal of Food Microbiology 66(1-2): 71-78.

Kistemann T, Classen T, Koch C, Dangendorf F, Fischeder R, Gebel J, Vacata V, Exner

M. 2002. Microbial load of water reservior tributaries during extreme rainfall and runoff.
App and Environ. Microbio. 68(5):2188-97.

114

Komecki TS, Sabbagh GJ, Storm DE. 1999. Evaluation of runoff, erosion, and
phosphorus modeling system- SIMPLE. J. Am. Water Res. Ass. 35(4): 807-820.

Liao HH, Tim US. 1997. An interactive modeling environment for non-point pollution
control. J. Am. Water Res. Ass. 33(3): 591-603.

Licence, K.; Oates, K. R.; Synge, B. A.; Reid, T. M. S. 2001. An outbreak of E-coli 0157
infection with evidence of spread from animals to man through contamination of a
private water supply. Epidemiology and Infection, 126(1) 135-138.

Mallin MA, Williams KE, Esham EC, Lowe RP. 2000. Effect of human development of
bacteriological water quality in coastal watersheds. Ecological Applications 10(4): 1047-
1056.

Mallin MA, Ensign SH, McIver MR, Shank GC, Fowler PK. 2001. Demographic,
landscape, and meteorlogical factors controlling microbial pollution of coastal waters.
Hydrobiologia 460: 1 85-193 .

Murray KS, Fisher LE, Therrien J, George B, Gillespie J. 2001. Assessment and use of
indicator bacteria to determine sources of pollution to an urban river. Journal of Great
Lakes Research 27(2): 220-229.

Michel P, Wilson JB, Martin SW, Clarke RC, McEwen SA. 1999. Temporal and
geographical distributions of reported cases of E. coli 0157:H7 infection in Ontario.
Epdemiol. Infect. 122: 191-200.

Noble RT, Weisberg SB, Leecaster MK, McGee CD, Dorsey JH, Vainik P, Orozco-
Borbon V. 2003. Storm effects on regional beach quality along southern California
shoreline. J. Water Health 01:1 23-31.

Niemi RM, Niemi J S. 1991 . Bacterial pollution of waters in pristine and agricultural
lands. J. Environ. Qual. 20:620-627.

Pettibone GW, Irvine KN. 1996. Levels and sources of indicator bacteria associated with
the Buffalo River area of concern. J of Great Lakes Research 22(4): 896-905.

Robinson KJ, Ragan RM. 1993. Geographic Information System based non-point
pollution modeling. Wat. Res. Bull. 29(6): 1003-1008.

Rose J B, Epstein PR, Lipp EK, Sherman BH, Bernard SM, Patz JA. 2001. Climate
variability and change in the Untied States: Potential impacts on water and foodbome
diseases caused by microbiologic agents. Environmental Health Perspectives 109: 211-
21 IS.

Rudra RP, Dickinson WT, Abedini MJ, Wall GJ. 1999. A multi-tier approach for
agricultural watershed management. J. Am. Water Res. Ass. 35(5): 1159-1170.

”5

Sample DJ, Heaney JP, Wright LT, Koustas R. 2001 Geographic information systems,
decision support systems, and urban storm water management. J. Water Res. Plan. and
Manag. ASCE. 127(3): 155-161.

Schepel M. 1998. STREAM: spatial tools for river basin, environment, and analysis of
mangement options, GIS. GIS Technologies and their Environmental Applications WIT
Press Southampton, Boston.

Shamsi UM. 1996. Storm water management implementation through modeling and GIS.
J. Water Res. Plan. and Manag. ASCE. 122(2): 114-127.

Smith JH, Wickman JD, Norton D, Wade TG, Jones KB, 2001, Utilization of landscape
indicators to model potential pathogen impaired waters. Journal of AWRA 37(4): 805-
814.

Stark SL, Nuckols JR, Rada J. 1999. Using GIS to investigate septic systems sites and
nitrate pollution. J. Envior Health, 61(8): 15-6.

Swerdlow DL, Woodruff BA, Brady RC, Grifiin PM, Tippen S, Donnell HD, E.
Geldreich E, Payne BJ, Meyer A, Wells JG, Greene KD, Bright M, Bean NH, and Blake
PA. 1992. A waterborne outbreak in Missouri of Escherichia coli 0157:H7 associated
with bloody diarrhea and death. Ann. Intern. Med. 117:812-819.

Tim US, Jolly R. 1994. Evaluating agricultural Non-point source pollution using
integrated geographical information systems and hydrologic/water quality model. J.
Envir. Qual. 23(1): 25-35.

USEPA. 1985. Test methods for Escherichia coli and enterococci in water by the
membrane filter procedure. Environmental Monitoring and Support Laboratory,
Cincinnati, OH. EPA-600/4-85/076.

USEPA. 1986. Ambient water quality criteria for bacteria-1986. Office of Water
Regulations and Standards, Criteria and Standards Division, Washington, DC. EPA-
440/5- 84/002.

USDA. 2002. AGNPS website http://www-.mrsursusda.gov/morris/productsx’zgnpshtm;
htlp://www.sedlabolemiss.edu;’agnps.html.

USEPA. 2002b. Better assessment science integrating point and non-point sources
(BASINS) http://wwwepa. gov/warterseiencef’iip-"basins/systcm/ BASIN S3/bit.htm

USEPA. 2003. Bacterial water quality standards for recreational waters (freshwater and
marine waters) EPA-823-R-O3-008.

116

Van Every LR. 1995. Groundwater as a vehicle for disease transmission in southeastern
Idaho: 3 case study. Journal of Environmental Health 58: 16-19.

Vitousek PM, Mooney HA, Lubchenco J, Melillo J M. 1997. Human domination of
earth’s ecosystems. Science 277: 494-499.

Wang XH, White-Hull C, Dyer S, Yang Y. 2000. GIS-ROUT: a river for watershed
planning. Envir. And Planning 27(2): 231-246.

Wang XH, Yin ZY. 1997. Using GIS to assess relationship between land-use and water
quality at a watershed level. Enviro. Intern. 23(1): 103-114.

Wang X. 2001. Integrating water quality management and land-use planning in a
watershed context. J. Enviro. Manag. 61(1): 25-36.

Wong KM, Strecker EW, Strenstrom MK. 1997. GIS to estimate storm-water pollutant
mass loadings. J. Envir. Eng. ASCE 123(8): 737-745.

117

Qflme—ré

A Linkage between Technology and Policy in Evaluating the Risk of Microbial

Safety in Recreational-use Water

Introduction

Ever since fecal contamination of water was determined as a human health risk,
there has always been a great deal of concern regarding the level of coliform bacteria
counts in water. Many bodies of water throughout the world are considered to have
bacteria counts above acceptable levels. The sources of these bacteria are thought to be
fecal contamination from humans, domestic animals and wildlife, as well as runoff from
agricultural land, inadequate septic systems or sewer overflow.

The United States Public Health Services (U SPHS) conducted a series of studies
in the 19403 and 19505 that showed a causal connection between fecal coliform and
gastrointestinal sickness exists (Stevenson 1953). In the late 19705, the EPA conducted a
landmark prospective cohort study and reported a linear relationship between the
incidence of gastroenteritis among swimmers and marine bacterial counts (Cabelli 1982).
Between 1973 and 1978, participants were recruited at beaches from three US. locations
(New York; Lake Pontchartrain, Louisiana; Boston, Massachusetts) and were contacted
by telephone days after going to the beach. Swimming status was self-selected, not
randomly assigned, and symptoms were self-reported. The mean proportion of swimmers
with gastrointestinal symptoms was 6.8% versus 4.6% in non-swimmers (Cabelli 1982).
When bacterial concentrations were above 1/100 ml, relative risk increased linearly,

reaching 4.0 with concentrations of 1,000/ 100 ml (p < 0.001). In addition, the frequency

118

of gastrointestinal symptoms was inversely related to the distance from known sources of
municipal wastewater (Cabelli 1982). Since these studies, a large body of literature
involving the epidemiology of illness from the use of recreational water has been
published (Examples include studies by: Seyfiied 1985; Van Asperen 1998; Henrickson
2001; Prieto 2001)

Beach closings and illness after exposure to marine water may be increasing in
frequency (Harvell 1991). There were over 12,000 coastal beach closings and advisories
in the United States from 1988 to 1994 (an increase of 400% over that period), with over
75% of the closings due to microbial contamination (Barton 1995). A recent report
compiled by the National Resources Defense Council (NRDC), which surveyed more
than 200 waterfront communities, found that during 1999 there were at least 6,160 beach-
use days of closings and advisories at beaches (N RDC 2000). According to the most
current report, the NRDC's twelfth annual beach report, at least 13,410 closings and
advisories were issued across the country in 2001, a 19 percent jump over the previous

year (NRDC 2002).

Public Policy to Reduce Risk in Recreational Water

One of the major goals of the Clean Water Act of 1972 and its amendments
(CWA) is to ensure that US. waters are safe for fishing and swimming (33 U.S.C.A. §§
1251 to 1387). The use of water quality indicators that accurately reflect "safety" is
essential for this goal. The CWA mandates the use of indicator organisms to determine
safe levels and CWA section 303(d) requires the reporting of “impaired waters.” In

addition, section 303, establishes the water quality standards and Total Maximum Daily

119

Load (TMDL) programs. The TMDL program is required by the rules that amended the

CWA issued in 1985 and then subsequently amended in 1992 (U SEPA 2002).

TMDL is a calculation of the maximum amount of a pollutant that a waterbody
can receive while still meeting water quality standards, and then determines an allocation
of that amount to the pollutant's sources. Water quality standards are set by States,
Territories, and Tribes and must meet or exceed the Federal standards set forth in the
CWA. These entities identify the uses for each waterbody, for example, drinking water
supply, contact recreation (swimming), and aquatic life support (fishing), and the
scientific criteria to support that use. A TMDL is the sum of the allowable loads of a
single pollutant from all contributing point and non-point sources. The calculation must
include a margin of safety to ensure that the waterbody can be used for the purposes the
State has designated. Additionally, the calculation must also account for seasonal

variation in water quality.

The indicator organism mandated for the monitoring of pathogens in fresh water
is E. coli. If the level of this organism is consistently elevated from the standard, then the
waterbody is declared an “impaired water” as defined in CWA 303 (d) and a TMDL is
required to be developed and implemented for that waterbody. As of 2001, there are
5,512 reported “impaired waters” due to pathogens (over 13% of all impairments
reported and this is the second highest impairment behind sediment which is just below

14%) (USEPA 2001).

The Beach Environmental Assessment Closure and Health (BEACH) Act was

signed into law in October 2000 and amended the CWA (U SEPA 2002b). This

120

legislation directs the EPA Administrator to conduct studies to review health risks; to
develop indicators to detect the presence of pathogens; to offer guidance for state-to-state
application of the revised water quality criteria; to publish and revise regulations
requiring the monitoring of coastal waters; to provide technical assistance to states for
uniform assessment and monitoring procedures; and to establish a national coastal
recreation water pollution occurrence database and a listing of communities complying
with regulations published pursuant to this legislation. In addition, the Beach Act
authorizes $30 million annually for grants that would help states, local governments and
Indian tribes to monitor beach waters and notify the public when beach water exceeds the
established criteria. Moreover, some of these grants are designated for the development

of improved detection of pathogens in water, both freshwater and marine water.

It is mandated that consistent national health standards for beach water be
established by 2004. Currently, the EPA has determined that E. coli level is the best
indicator of risk to swimmers in fresh water. An E. coli concentration of less then 126
CF U/ 100ml, calculated as a geometric mean over 30 days, is considered to be safe for
full body contact recreational use of water (U SEPA 1986). If a waterbody has a
designated use of full body contact recreational use and the E. coli levels are consistently
elevated from the above described standard, then the waterbody is declared an “impaired
water” as defined in CWA 303 (d) and a TMDL is required to be developed and

implemented for that waterbody.

Many county health departments in the State of Michigan routinely collect water

samples at beaches to determine if the water is safe for swimming. Samples are generally

121

taken one foot below the surface in water that is between three and six feet in depth. The
analysis is performed in a laboratory using standard methods. E. coli bacteria are counted
and judged against standards established by state rules. Results from the method are
available after approximately 28 hours; so the results from this analysis are reported the
following day.

County health departments take a minimum of three samples each time a beach
area is monitored. The dfly geometric mean calculated from these samples must be
below 300 E. coli per 100 milliliters for the water to be considered safe for swimming.
One or two of the samples may be above 300, but if the daily geometric mean is below
300, the beach is not in violation of the water quality standard. A minimum of five
sampling events (consisting of at least three samples per event) must be collected within a
30-day period for the results to be considered a reliable indication of water quality. After
30 days, a geometric mean is calculated for all the individual samples collected within
that time frame. This My geometric mean must be below 130 E. coli per 100 ml for
the water to be considered safe for swimming. Although the method yields results that
are very accurate, the results are not reported in less that 24 hrs. Officials close beaches
based on old information that may or may not be the actual condition of the water at the

time of closure and leave prior groups exposed.

E.coli 0157:H7

In the early19808, E. coli 0157:H7 was recognized as a pathogen (Riley 1982).
The toxicity to humans from this pathogen has been reported to be as low as 10 cells

(Phillips 1999). There are an estimated 73,000 cases of E. coli 0157 infections per year

122

in the United States, of which approximately 62,000 are food-home and 11,000 are
waterborne (Mead 1999). These estimates consider waterborne cases fi'om ingestion of
water as food not from recreational use of a waterbody. Examples of waterborne cases
include Walkerton, Ontario where the town’s water supply was contaminated by E. coli
0157:H7 (not included in the estimations because it happened in Canada in the year
2000). The identified cases of E. coli 0157 infections from the use of recreational water
are limited to a few epidemiological studies (Mudgett 1998; Friedman 1999; Olsen

2002)

Since the current accepted analysis for E. coli and E. coli 0157:H7 takes days,
officials close beaches based old information that may or may not be the actual condition
of the water at the time of the closure and thus leaving prior groups exposed. What is
needed is a rapid method that can give information on the beach conditions so that
officials can make decisions on closure before the public is exposed to the water. This is
supported by a study which conducted on beaches in Lake Michigan over a 12-year
period concluded that current protocol is inadequate for predicting risk at the time of
beach closure (Whitman 1999). The biosensor in this study that is evaluated and
compared to the existing methods produces results in about 10 minutes and is
inexpensive. It is believed that the use of such a device would increase the public’s

safety in the use of recreational water.

Biosensors
A biosensor is an analytical device that integrates biological sensing elements

with a transducer (Turner 1998; Ivnitski 1999). The general function of a biosensor is to

123

convert biological events into a quantifying electrical response (Cahn 1993).
Electrochemical irnmunoassays are biosensors made up of antibodies as biological
sensing elements attached to an electrochemical transducer (Zhang 2000). The
performance of a biosensor based on polyaniline as an electrochemical transducer in
measuring an immunological reaction has been evaluated (Muhammad-Tahir and
Alocilja, 2002). Polyaniline in particular, has been one of the most extensively
investigated conducting polymers, due to its excellent stability in liquid form, promising

electronic properties (Syed 1991) and strong biomolecular interactions (Irnisides 1996).

Objectives

The main focus of this project is to improve the safety of recreational-use water
through better communication of the actual risk. The specific objectives are: l) to
determine if rain or river flow has an influence on the concentration of E. coli and E. coli
0157:H7; 2) to evaluate if total E. coli measured maybe used as an indicator of
pathogenic E. coli 0157:H7 contamination; 3) to evaluate if a biosensor maybe employed

to measure quickly the presence of E. coli and E. coli 0157:H7.

Materials and Methods
The Proiect Site

The Ingham County Health Department (ICHD), Michigan Department of
Environmental Quality (MDEQ), Michigan Department of Community Health (MDCH),
and Michigan State University (MSU) collaborated on a study to determine the amount of

E. coli and pathogenic E. coli 0157:H7 in the Red Cedar River. For four weeks in April

124

and May of 2002, four sites were sampled (Hagadom, Farm Lane, Kalamazoo Street, and
Putman Street; See Figure 5-1).

The samples were split with one going to MDEQ for total E. coli analysis and
then the cultures were given to MDCH for the determination of E. coli 0157:H7. The
Biosensor Lab in the Agricultural Engineering Department at MSU analyzed the other
split sample. The Biosensor Lab used a recently invented biosensor built specifically for

the rapid analysis E. coli 0157:H7.

   

Farm Lane

Hagadom

Figure 5-1. Red Cedar River Watershed with Locations of Sampling Sites
The Red Cedar River arises in Cedar Lake in the south-central Lower peninsula of
Michigan and flows about 45 miles (80 kilometers) to its confluence with the Grand

River in the city of Lansing. It has 12 tributaries and drains a total area of about 472

125

square miles (22,000 hectares). The river provides mid-Michigan residents with
numerous recreational opportunities that include angling, canoeing, kayaking,
photography and bird watching. The river also serves as a source of water for the
irrigation of crops throughout the watershed. In terms of aesthetics, hundreds of
thousands enjoy a walk along the Red Cedar River on the campus of MSU each year.
Three of the weekly samples (Hagadom, Farm Lane, and Kalamazoo Street) were taken
from the 5,400 acre MSU campus which has agricultural operations two miles south of
the river, traditional campus setting along the river, and the urban landscape of the city of
East Lansing less than a quarter mile north of the river. The fourth weekly sample was
taken from the Putrnan Street Bridge that is surrounded by the city of Williamston,
Michigan near a public kayak course and is approximately 20 miles up stream from the
MSU campus. The Red Cedar River has been identified as being impaired as defined in
the CWA section 303(d) for pathogens and dissolved oxygen (U SEPA 2000). It is
anticipated that a TMDL will be implemented by December 31, 201 1 for both
impairments.

Lowering a specially designed bottle holder with a 100ml collection bottle into
the Red Cedar River made it possible for the sample to be collected. The bottle holder
allowed the technician to take the sample from any bridge spanning the Red Cedar. The
sampling bottle was lowered fi'om the center of the bridge until it was approximately two
feet below the surface of the river and then retrieved full of river water. Four bottles
were collected at each site. After the collected samples were properly labeled, they were

put in a cooler filled with ice. Three of the four 100ml samples were dropped off to the

126

MSU biosensor lab for analysis and the remaining 100ml sample for each site was

delivered to the MDEQ for analysis.

Weather Rel_ated Data Collection

Along with the variation of the concentration of E. coli during the study, there
were variations in rainfall and river flow for the different sampling events. The National
Weather Service (N WS) reported rainfall from a station in Williamston, Michigan where
the Putrnan Street site is located. The Horticulture Research Station reported rainfall for
the MSU campus on the campus. The United States Geological Services (U SGS)
provided the river flow data from USGS staging stations in Williamston and on the MSU

campus.

E. coli Materials and Method_s (EPA Method 1103.1)

The original mTEC Agar enumeration method (Dufour et al., 1981) for E. coli
was introduced by EPA in 1986 (U SEPA, 1986b). A revised method was developed in
1998 by the EPA and has been designated as the modified mTEC method. Both the
mTEC and modified mTEC Agar methods use the membrane filter procedure. The two
membrane filter methods provide a direct count of E. coli in water based on the
development of colonies that grow on the surface of the membrane filter (Difco Labs,
Detroit MI). The MDEQ used the revised method for the analysis in this study.

A water sample is filtered through the membrane, which retains the bacteria. After
filtration, the membrane containing the bacteria is placed on a selective and differential

modified mTEC Agar medium, incubated at 352t0.5°C for 2 h to resuscitate the injured or

127

stressed bacteria, and then incubated at 44.5:t0.2°C for 22 h. After 20 minutes, Red

colonies on modified mTEC medium are counted with the aid of a fluorescent lamp and a
glass lens (2—5x magnification) or stereoscopic microscope. The Enumerated colonies are
forwarded to MDCH for E. coli 0157:H7 biochemical confirmation and serotyping. This

procedure adds a few more days before results are finally reported.

Biosenflr Material and Methods

The biosensor used in this analysis consists of two parts: The immunosensor and
the electronic data collection system. The immunosensor is comprised of four different
pads: sample application, conjugate, detection, and absorption. The system was
constructed as shown in Figure 5-2. The cellulose membrane was used (5 x 10 mm) for
the sample application pad, fiber glass membrane for the conjugate pad (5 x 10mm),
nitrocellulose (NC) membrane (5 x 20 mm) for the capture pad, and cellulose membrane
(5 x 30 mm) for the absorption pad. Silver electrodes were fabricated on the NC
membrane to electrically connect the immunosensor with the electronic data collection

system consisting of a copper wafer and an ohmmeter linked to a computer.

128

Analyte

(A) Absorptlon
Conjugate membrane
Membrane

I

Electrode ."

   

Application ¢ Wafer
mem bra ne I
Captu re

membrane

Electrode

Figure 5-2. Schematic diagram of the immunosensor. Conjugate pad for polyaniline-
labeled antibody absorption (A). Detection pad coated on each side with silver electrodes
(B). Gap between electrodes is the site for antibody immobilization

Aniline, glutaraldehyde, N, N Dimethylformamide (DMF), Tween-20, lithium
chloride, tris buffer, phosphate buffer saline (PBS) were purchased from Sigma-Aldrich
(Missouri). Anti- E. coli 0157:H7 was obtained from Kirkegaard & Perry Laboratories,
Inc. (Gaithersburg, MD). Nitrocellulose (NC) membrane with a flow rate of 160 sec per 4
cm, cellulose membrane, and fiberglass membrane grade G6 were purchased from
Millipore (Massachusetts). Micro-tip silver pen was purchased from Chemotronics
(Georgia). Other reagents used were of analytical grade. All chemicals and diluents were
prepared with doubly deionized water with conductivity below 0.1 uS/cm. Lyophilized
affinity purified polyclonal antibodies of E. coli 0157:H7 were purchased, and stored at
4°C until rehydrated with phosphate buffer (pH 7.0). Two different concentrations of
antibodies were prepared: 50011ng of antibody was used on the capture pad, and
150ug/m1 was used for conjugating with polyaniline and coated on the conjugate pad.

A water-soluble polyaniline was synthesized by following a standard procedure of

oxidative polymerization of aniline monomer in the presence of ammonium persulfate

129

(Kim, 2000). A mixture of the antibody and polyaniline was left to react for 30 minutes.
The conjugate was then precipitated by centrifugation (13000 rpm for 3 min) using 0.1M
Tris buffer as the blocking reagent. The conjugated antibody was diluted in 0.01M LiCl
and stored at 4°C before use. When ready, lOul of the conjugated antibody was applied
on the conjugate pad and left to dry. Cellulose membranes for absorption and sample
application pads were treated with distilled water three times to remove dirt and surface
residuals. The membranes were left to dry and stored inside a petri dish to maintain a
clean surface.

Affinity purified antibodies were immobilized on the NC membrane by the
following steps. First, the NC membrane was saturated in 10% (v/v) methanol in water
for 45 min and left to dry. The membrane was then treated in 0.5% (v/v) glutaraldehyde
for 1 hour. After drying, 2.5ul of 0.5mg/ml of antibody was pipetted on the membrane,
and incubated at 37°C for one hour. Inactivation of residual fimctional groups and
blocking was carried out simultaneously by incubating the membrane with 0.1M tris
buffer, pH 7.6, containing 0.1% Tween-20 for 45 min. The membrane was left in the air
to dry.

All prepared membranes were arranged in the order mentioned in Figure 2 and
attached onto an etched copper plate using double-sided tape. The prepared biosensor
was stored at 4°C before its use. The prepared biosensor was connected to a multimeter
and a computer for signal measurement. To begin detection, 0.1 ml of the water sample
was pipetted onto the application pad. The generated signal was captured using a
multimeter (BK multimeter, MA) in the form of resistance. The biosensor was calibrated

using the uninoculated distilled water.

130

Two tests were performed in this study to assess the performance of the biosensor.
Sensitivity testing was conducted to evaluate the detection limit of the biosensor. In this
testing, a distilled water sample inoculated with a varying concentration of E. coli
0157:H7 was used.

The second test involved with the water samples collected from various sources
along the Red Cedar River: Kalamazoo Street, Farm Lane, Hagadom, and Putrnan Street.
One hundred ml of each sample was filtered through a bio-filter membrane (BIOPath,
Florida). The membrane filter was transferred to a test tube containing 10 ml of 0. l %
peptone water. The test tube was vortexed to release the trapped cells from the
membrane. Then, 0.1 ml of each water sample was pipetted onto the application pad of

the biosensor. Three replicates were done for each treatment.

Results and Discussion

Pertinent data collected from the four sampling sites shown in Figure 1 are listed
in the Table 5-1. The data in Table 5-1 includes E. coli concentration in colony forming
units (CF U) per 100ml as reported by the MDEQ and E. coli 0157:H7 concentration in
CF U per 100ml as reported by the MDCH. In addition, rainfall data that was calculated
in the 24-hour period, 48-hour period, and 72-hour period prior to the collection of the
water samples that was analyzed for the E. coli and E. coli 0157:H7 concentrations is
included in Table 5-1. The Red Cedar River flow rate in cubic feet per second of water

moved is also reported in Table 5-1 for the dates of sampling.

131

Table 5-1. E. coli Concentration, Rainfall, River Flow, and E. coli 0157:H7
Concentration from the Sampling Sites in the Red Cedar River Watershed.

Kalamazoo Street

Date E.coli Rain 24 hrs. Rain 48 hrs. Rain 72 hrs. Flow 0157:H7
(CPU/100ml) (inches) (inches) (inches) (ft3/sec) (CPU/100ml)
4/15/2002 190 0 0 0.03 387 0
4/22/2002 60 0.15 0.15 0.17 232 0
4/30/2002 250 0 0.2 0.53 216 0
5/07/2002 110 0.17 0.17 0.17 229 0
5/14/2002 720 0.03 0.2 1.09 456 0
Farm Lane
Date E.coli Rain 24 hrs. Rain 48 hrs. Rain 72 hrs. Flow 0157:H7
(CF U/ 100ml) (inches) (inches) (inches) (ft3/sec) (CFU/100ml)
4/15/2002 173 0 0 0.03 387 0
4/22/2002 47 0.15 0.15 0.17 232 0
5/07/2002 107 0.17 0.17 0.17 229 0
5/14/2002 1 150 0.03 0.2 1.09 456 0
Hagadorn
Date E.coli Rain 24 hrs. Rain 48 hrs. Rain 72 hrs. Flow 0157:H7
(CPU/100ml) (inches) (inches) (inches) (ft3/sec) (CF U/ 100ml)
4/15/2002 90 0 0 0.03 387 0
4/22/2002 80 0.15 0.15 0.17 232 0
5/07/2002 130 0.17 0.17 0.17 229 0
5/14/2002 770 0.03 0.2 1.09 456 0

Putman Street

Date E.coli Rain 24 hrs. Rain48 hrs. Rain 72 hrs. Flow 0157:H7
(CPU/100ml) (inches) (inches) (inches) (ft3/sec) (CPU/100ml)

4/16/2002 43 0 0.1 0.1 197 0
4/30/2002 77 0 0.14 0.41 86 0
5/07/2002 133 0.06 0.06 0.06 87 0
5/14/2002 453 0.05 0.13 0.87 227 0

132

Objective 1: Rain aLd River Flow influence on E. coli Concentration

Along with the variation of the concentration of E. coli during the study, there
were variations in rainfall and river flow for the different sampling events. The National
Weather Service (NWS) reported rainfall from a station in Williamston, Michigan where
the Putrnan Street site is located. The Horticulture Research Station reported rainfall for
the MSU campus on the campus. The United States Geological Services (USGS)
provided the river flow data from USGS staging stations in Williamston and on the MSU
campus. This flow data is provisional at the time of calculations and will not be reviewed
and made official until the spring of 2004.

In the statistical analysis that was performed and shown in Table 5-2, rainfall was
collected for the 24 hour period of the day of the sample collection and in addition the
cumulative total of rainfall for 48 hour period and 72 hour period. The statistical analysis
indicates that there is an influence of higher concentration of E. coli from both increased
rainfall and increased river flow. There are older studies that reported similar findings
(Stephenson 1978; Tiedemann 1988). More recently, in a three-year study in Arkansas,
it was reported that E. coli concentrations increased with increasing flow rates (Edwards
1997). In addition, a study that examined 99 samples from three tributaries that
contributed to different drinking water reservoirs showed that E. coli levels along with
other bacteriological parameters increased considerably during extreme rainfall events
(Kistemann 2002). It is interesting to note that the 24 hour period of rainfall is not
statistically significant in influencing E. coli concentration but this may be explained by
the fact that the rainfall total that is reported is from midnight to midnight of a day and

the sampling was done in the late morning. Since the rainfall collection includes time

133

that is after the sampling is done, it might correctly reflect the rainfall impact in the 24-
hour period. In both the Pearson’s correlation and ANOVA the total rainfall for a 72-

hour period was statistically significant at p < 0.0001.

Table 5—2. Statistical Analysis of E. coli Concentration vs. Rainfall, River Flow, and
E. coli 0157:H7 Concentration.

Statistical Formula Rain 24 hrs Rain 48 hrs. Rain 72 hrs. Flow 0157:H7
Pearson’s Correlation -.315 .429* .899” .674“ NA
ANOVA .082 .267 14.129" 4.760* NA

* Statistically significant at p < 0.05; ** Statistically significant at p < 0.01.
NA = Statistical analysis not possible since all entries were zero.
Obiective 2: E. coli a_s_ an Indicator of Pathogenic Contamination

The E. coli concentrations varied from a very low level of 43 CFU/100ml to a
very high level of 1150 CFU/100ml, however, no E. coli 0157:H7 was detected in any of
the samples analyzed. Within the parameters of this study, the results show that the use
of total E. coli concentration as an indicator of E. coli 0157:H7 does not provide a
reliable risk factor of the pathogenic bacteria to humans using the water. The statistical
analysis shown in Table 5-2 indicates that there is no correlation between E. coli and E.
coli 0157:H7 concentrations.

This suggests that E. coli levels may not be an accurate indicator of the level of
pathogenic bacteria in the water. Therefore the use of E. coli levels for determining the
risk of pathogenic E. coli 0157:H7 to humans using recreational water may have to
redefined. A study in Italy concluded that concentrations of indicators of fecal
contamination, whether within or over the values established by legislation, are not

always related to the presence and density of individual pathogenic serotypes

134

(Bonadonna 2002). In this reported study, for many cases, the values of pathogenic

serotype levels were not linear with respect to the levels of the bacterial indicators.

Obiective 3: Biosensor Performance

Figure 5-3 illustrates the response of the biosensor tested with inoculated distilled
water samples. The resistance drop was determined by calculating the difference between
the resistance outputs of the blank and that of the inoculated samples. The difference in
measured resistance (resistance drop) was the reduction in resistance due to electron
transfer facilitated by the polyaniline-labeled antibody between the electrodes.

In water samples inoculated with E. coli 0157:H7, the signal was shown to be
proportional to the cell concentration between 101 and 104 CFU/ml. The signal was
observed to decrease at concentrations higher than 104 CF U/ml (Figure 5-3). Statistical

analysis reveals that the sensitivity limit of the biosensor was at 8.3 d: 0.1 X 10' CF U/ml.

135

 

 

 

50.00

40.00

30.00 -

20.00

Redefines drop (Kohm)

10.00

 

0.00

 

10*3 10*4 10*5
Coll concontratlons (CPU/ml)

 

 

 

Figure 5-3. Resistance drop of the biosensor tested in distilled water samples inoculated
with E. coli 0157:H7.

The biosensor did not report any false positive during this study and the results were
identical to the reported concentrations fi'om traditional methods that use the culturing of
the pathogenic bacteria as performed by MDCH shown in Table 3. The levels of E. coli
0157:H7 remained zero despite a wide variation in the levels of total E. coli. The
advantage in the biosensor is that it produced results in less than 20 minutes in contrast to

the traditional method that produced results in three days.

136

Table 5-3. E. coli 015 7:H7 Analyzed by Traditional Methods and by the Biosensor.

Kalamazoo Street

Date E. coli

(CF U/ l 00ml)
4/15/2002 190
4/22/2002 60
4/30/2002 250

5/07/2002 1 10

5/14/2002 720
Farm Lane
Date E. coli
(CFU/ 100ml)
4/15/2002 173
4/22/2002 47
5/07/2002 107
5/14/2002 1150
Hagadorn
Date E. coli
(CFU/100ml)
4/15/2002 90
4/22/2002 80
5/07/2002 130
5/14/2002 770
Putman Street
Date E. coli
(CPU/100ml)
4/16/2002 43
4/30/2002 77
5/07/2002 133
5/14/2002 453

E. coli 0157:H7 (MDCH)
(CPU/100ml)

OCOCO

E. coli 0157:H7 (MDCH)
(CPU/100ml)

COCO

E. coli 0157:H7 (MDCH)
(CPU/100ml)

COCO

E. coli 0157:H7 (MDCH)
(CPU/100ml)

COCO

E. coli 0157:H7(Biosensor)
(CF U/ 1 00ml)

OCOCO

E. coli 0157:H7 (Biosensor)
(CF U/ 100ml)
0

0
0
0

E. coli 0157:H7 (Biosensor)
(CPU/100ml)

COCO

E. coli 0157:H7 (Biosensor)
(CPU/100ml)

COCO

Additional studies are needed to determine the repeatability and reliability of this

device. Additional pathogenic serotypes, as well as, total E. coli could be adapted to the

biosensor and studies need to be conducted to test such configurations. The biosensor

137

was only tested in freshwater and results from marine (salt water) conditions are
necessary to determine the versatility of this device. Finally, since this study was
conducted on a river, a similar study needs to be conducted for beaches located on lakes.
With additional studies and implementing the findings from these future studies, in the
near future the biosensor may provide the best available method for determining the
safety of recreational-use water. Image this: a lifeguard dips a biosensor type device into
the lake water. This device tests the water for the levels of many different strains of
bacteria. Using the biosensor device is as easy as testing for pH. The results are
available in minutes and the lifeguard can determine if the water is safe before the beach

is opened to the public.

Conclusions

Methods of measuring pathogens directly are costly and time-consuming.
Therefore, indicator organisms such as E. coli are used instead of analyzing the
pathogens themselves. In many cases the TMDL implementation plan specifies the use
of best management practices (BMPs) or a systems of BMPs. Water quality monitoring
is an important component of the TMDL implementation plan and is needed to measure
the success of both individual activities and the overall water quality of the waterbody.
Although the samples were limited, the biosensor has shown promise for water quality
analysis both before and after the installation of the BMPs. Since the biosensor analysis
is both rapid and inexpensive, it could be an excellent tool for determining the success of
installed BMPs non point pathogenic pollution as well as monitoring the TMDL

implementation plan’s improvement of the water quality.

138

The 015 7:H7 results were identical for all samples by both the standard methods
and the biosensor. The biosensor did not report any false positive. E. coli 0157:H7 was
zero despite a wide variation in the levels of total E. coli. This suggests that E. coli levels
may not be an accurate indicator of the level of pathogenic bacteria in the water and
therefore the use E. coli levels for determining the risk to humans using recreational
water may suggest a risk that is greater than the true risk for the use of the water. More
data is needed to confirm these findings and also to compare levels of other strains of

bacteria to total E. coli, which is the defined indicator bacterium for water quality.

References

Barton K, Fuller D. 1995. Testing the Waters V: Politics and Pollution at US. Beaches.
New YorkzNatural Resources Defense Council.

Bonadonna L, Briancesso R, Coccia AM, Semproni M, Stewardson D. 2002. Occurrence
of Potential Bacterial Pathogens in Coastal Areas of the Adriatic Sea. Environ. Monitor
and Assess. 77(1):31-49.

Brunelle, S. 2001. Electroirnmunoassay technology for F oodbome pathogen detection.
Detection Technology: 55-62.

Cabelli VJ, Dufour AP, McCabe LJ, Levin MA. 1982. Swimming-associated
gastroenteritis and water quality. Am J Epidemiol 115:606-616.

Cahn, T. M. 1993. Biosensors. London, UK, Chapman and Hall.
D'Souza, S. F. 2001. Microbial Biosensors (Review). Biosensor & Bioelectronics 16:
337-353.

Dufour AP, Strickland ER, Cabelli VJ. 1981. Membrane Filter Method for Enumerating
Escherichia coli. Appl Environ Microbiol 41 :1 152-1158.

139

Edwards DR, Coyne MS, Daniel TC, Vendrall PF, Murdoch JF, Moore JF. 1997.
Indicator Bacteria Concentration of Two Northwest Arkansas Streams in Relation to
Flow and Season. Transactions of ASAE 40 (1):103-109.

Friedman, M. S., T. Roels, et al. 1999. Escherichia coli 0157 : H7 outbreak associated
with an improperly chlorinated swimming pool. Clinical Infectious Diseases 29(2): 298-
303.

Harvell CD, Kim L, Burkholder JM, Colwell RR, Epstein PR, Grimes DJ, Hofrnann EE,
Lipp EK, Osterhaus AD, Overstreet RM, et al., 1999. Emerging marine diseases-climate
links and anthropogenic factors. Science 285:1505-1510.

Henrickson, S. E., T. Wong, et al. 2001. Marine swimming-related illness: Implications
for monitoring and environmental policy. Environmental Health Perspectives 109(7):
645-650.

Hoa, D. T., Suresh Kumar, T.N., Punekar, N.S., Srinivasa, R.S., Lal, R., A.Q. 1992.
Contractor, Biosensor based on conducting polymers. Analytical Chemistry 64: 2645-
2646.

Ivnitski, D., Abdel-Hamid, I., Atanasov, P., Wilkins, E. 1999. Biosensor &
Bioelectronics 14: 599-624.

Kim, J. H., Cho, J .H., Cha, GS. 2000. Conductimetric membrane strip immunosensor
with polyaniline-bound gold colloids as signal generator. Biosensor & Bioelectronics 14:
907-915.

Kistemann T, Classen T, Koch C, Dangendorf F, F ischeder R, Gebel J, Vacata V, Exner
M, 2002. Microbial Load of Water Reservior Tributaries During Extreme Rainfall and
Runoff. App and Environ, Microbio. 68(5):2188-97.

Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin,
and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect.
Dis. 5:1-32

Muhammad-Tahir, Z. and Alocilja, E.C. Fabrication of a membrane strip immunosensor.
IEEE Sensors Journal, accepted for publication.

Mudgett, C. C., R. Ruden, et al. 1998. A beach-associated outbreak of Escherichia coli
0157 : H7. Journal of Environmental Health 60(9): 7-13.

NRDC, 2000, 11th Annual Survey of Water Quality Monitoring and Public
Notification Programs at US. Beaches.

NRDC, 2002, 12th Annual Survey of Water Quality Monitoring and Public
Notification Programs at US. Beaches. http://www.nrdc.org/water/oceans/ttw/titinx.asp

140

Olsen, S. J ., G. Miller, et al. 2002. A waterborne outbreak of Escherichia coli 0157 : H7
infections and hemolytic uremic syndrome: Implications for rural water systems.
Emerging Infectious Diseases 8(4): 370-375.

Phillips, C. A. 1999. The epidemiology, detection and control of Escherichia coli 0157.
Journal of the Science of Food and Agriculture 79(11): 1367-1381.

Prieto, M. D., B. Lopez, et al. 2001. Recreation in coastal waters: health risks associated
with bathing in sea water. Journal of Epidemiology and Community Health 55(6): 442-
447.

Seyfried, P. L., R. S. Tobin, et al. 1985. A Prospective-Study of Swimming-Related
Illness .1. Swimming— Associated Health Risk. American Journal of Public Health 75(9):
108-1070.

Seyfried, P. L., R. S. Tobin, et al. 1985. A Prospective-Study of Swimming-Related
Illness .2. Morbidity and the Microbiological Quality of Water. American Journal of
Public Health 75(9): 1071-1075.

Sergeyeva, T. A., Lavrik, N.V., Rachkov, A.E. 1996. Polyaniline label-based
Conductimetric Sensor for IgG detection. Sensors and Actuators B(34): 283-288.

Stephenson GR, Street LV.1978. Bacterial Variations in Streams from a Southwest Idaho
Rangeland Watershed. J. Environ. Qual. 7(1):]50-157.

Tiedemann AR, Higgens DA, Quigley TM, Sanderson HR, Bohn CC. 1988. Bacterial
Water Quality Responses to Four Grazinf Strategies. J. Environ. Qual. l7(3):492-498.

Turner, A. P., Newman, J .D .1998. An Introduction to Biosensor. Biosensor for Food
Analysis. UK, Athenaeurn Press Ltd: 13-27.

USEPA. 1985. Test methods for Escherichia coli and enterococci in water by the
membrane filter procedure. Environmental Monitoring and Support Laboratory,
Cincinnati, OH. EPA-600/4-85/076.

USEPA. 1986a. Ambient water quality criteria for bacteria-1986. Office of Water
Regulations and Standards, Criteria and Standards Division, Washington, DC. EPA-
440/5- 84/002.

USEPA. 1986b. Bacteriological ambient water quality criteria; availability. Federal
Register 51(45):8012-8016.

USEPA. 2000. State Impairments-Michigan, List ID M1082828H

USEPA. 2001. Protocol for Developing Pathogenic TMDLs, EPA 84] -R-00-002.

141

USEPA. 2002a. http://www.epa.gov/owow/tmdl/intro.html
USEPA. 2002b. http://www.epa.gov/waterscience/beaches/act.html

van Asperen, I. A., G. Medema, et al. 1998. Risk of gastroenteritis among triathletes in
relation to faecal pollution of fresh waters. International Journal of Epidemiology 27(2):
309-315.

Whitman RL, Nevers MB, Gerovac PJ. 1999. Interaction of Ambient Conditions and
Fecal Coliform Bacteria in Southern Lake Michigan Beach Waters: Monitoring Program
Implications, Natural Areas J. 19(2): 166-171.

Zhang, S, Wright, G., Yang, Y. 2000. Materials and techniques for electrochemical
biosensor design and construction. Biosensors & Bioelectronics 15:273-282.

142

Chapter 6

Conclusions and Recommendations

Ever since fecal contamination of water was determined a human health risk, there
has always been a great deal of concern regarding the level of coliform bacteria counts in
water. Many bodies of water throughout the world are considered to have counts above
acceptable levels. The sources or pathways of these E. coli are thought to be fecal
contamination from humans, domestic animals and wildlife, as well as runoff from
agricultural land, inadequate septic systems or sewer overflow. Physical factors, such as
seasonal variability, rainfall, river flow, nutrient and the survival of bacteria in water,
have an impact on E. coli levels in a waterbody. Tools exist to assist in determining the
risk to swimmers of bacteria illness in a watershed. Although risk assessment models
have been studied, an accurate dose-response model for illness to swimmers of bacteria
illness in a watershed is not available.

There is a large body of literature involving the epidemiology of illness from the
use of recreational water. Elevated E. coli levels have been identified as the cause of the
illness and E. coli levels have been mandated as the indicator bacteria for the
determination of water safety for swimmers. The validity of E. coli as an indicator of risk
of fecal contamination has been debated and E. coli may not be an accurate indicator of
pathogenic risk.

In contrast to the numerous studies, cases and reports of health effects from
exposure to indicator bacteria in recreational-use water, there are a limited number of

such findings in literature in regards to pathogenic E. coli 01 5 7:H 7. In all these reports

143

or studies that have identified E. coli 01 5 7:1-1 7 illness related to swimming, only one
identified E. coli 015 7:H7 in the water that was the exposure. Although E. coli 015 7:H7
infection from swimming is very rare, the reports indicate that children are at much
higher risk that adults. This higher risk may be from weaker immune systems of children
as compared to adults but more likely, it is due to the fact that children are more likely to
ingest water while swimming and that children are more likely than adults to defecate
while swimming. The use of scaled rubber swimming trunks on children maybe have
the greatest impact in lowering the risk of E. coli 015 7:H7 infections when swimming.

Public policy in both the United States and around the world has addressed the
risk to swimmers of illness from bacteria in water. The levels of indicator bacteria that is
acceptable as to the risk for swimmers have been established by USEPA, EEC, and
WHO. In addition to these policies, laws and guidelines have been established to
decrease the microbial input in a waterbody from sewage sources and agricultural
facilities. Many of these laws and guidelines are being reviewed and revised in order to
improve the public health of swimmers.

Based on the data collected and analyzed in this study, it is concluded that there is
no observed seasonal effect for E. coli levels from the Farm Lane Bridge over the three-
year period from 2000 to 2002. From the statistical analysis using Spearman’s rho, river
flow, rain, humidity, low temperature of the sampling day and soil moisture were
significantly (p<0.05) correlated to E. coli concentrations.

Other studies support the association of the levels of fecal indicators in a water
body and rainfall events and the results presented in this study support these findings. In

addition, previous studies have shown that increased river flow has been related to

144

increases in the concentration levels of E. coli in a river and the results presented in this
study also support these findings. From this study there is no correlation between E. coli
concentration and the number of ducks in the river upstream from the sampling point.
Solar radiation and wind speed did not correlate to E. coli concentrations.

Using governmental guidelines for maximum E. coli level for the safe use of
water for recreation, statistical models were designed and it was concluded that the odds
of E. coli exceeding 300 cfu/100ml are increasing at least twofold and at most twelve-fold
for each 1 cm increase in 72-hour total rainfall. It was also concluded that if the low
temperature of the sampling day was greater than 15°C degrees the 95% probability of
E. coli exceeding 300 cfu/100ml occurs at 0.25 cm 72-hour total rainfall.

Based on the results from this study, it is concluded that elevations of nutrients do
have an association with elevations in the levels of E. coli concentration. The
concentration levels of ammonia nitrate were significantly correlated to E. coli
concentrations and the levels of total suspended solids were significantly correlated to E.
coli concentrations.

These results indicate that if the locations of the sources of nutrients that enter
into nearby surface water are identified then proper best management practices (BMP)
can be implement at these locations and the implemented BMP will have an impact on
controlling nutrients into the surface water as well as having an effect to lower the E. coli
concentration in the surface water.

However, in this study rainfall did not show significant positive correlation with
E. coli levels and this is contradictory to what has been reported by others. This

contradiction maybe explained by the fact that the rainfall data was collected fi'om three

145

different sites that include East Lansing, Williamston, and Howell so that the rainfall data
better represents the sampling site. These rainfall collection sites were not at the 17
sampling sites and probably did not represent the real rainfall that occurred at the 17
sampling sites during the time period that the river water samples were collected.

The impact of land use on E. coli concentrations was investigated but the results
did not reveal with any certainty as to which type of land-use has the greatest impact on
elevated E. coli concentrations. In the small sample set that was analyzed, agricultural
land use had the greatest impact on elevated total suspended solids. However, mixed
land use had the greatest impact on elevated ammonia nitrates. Larger sample sets need
to be studied to determine if the type of land-use does have an impact the concentration
levels of E. coli and other nutrients in nearby surface water.

Methods of measuring pathogens directly are costly and time-consuming.
Therefore, indicator organisms such as E. coli are used instead of analyzing the
pathogens themselves. In many cases the TMDL implementation plan specifies the use
of BMPs or a systems of BMPs. Water quality monitoring is an important component of
the TMDL implementation plan and is needed to measure the success of both individual
activities and the overall water quality of the waterbody. The biosensor could be used
for water quality analysis both before and after the installation of the BMPs. Since the
biosensor analysis is both rapid and inexpensive, it could be an excellent tool for
determining the success of installed BMPs non point pathogenic pollution as well as
monitoring the TMDL implementation plan’s improvement of the water quality.

The E. coli 015 7:H7 results were identical for all samples by both the standard

methods and the biosensor. The biosensor did not report any false positives. The levels

146

of E. coli 015 7:H7 remained constant in relation to variation in the levels of total E. coli.
This suggests that E. coli levels may not be an accurate indicator of the level of
pathogenic bacteria in the water and therefore the use E. coli levels for determining the
risk to humans using recreational water may suggest a risk that is greater than the true
risk for the use of the water. Since this collaborative study was done with out funding, it
was only operational for a four-week period. More data is needed to confirm these
findings and also to compare levels of other strains of bacteria to total E. coli, which is
the defined indicator bacterium for water quality.

However, there are weaknesses of this study. The biosensor was only tested in
freshwater and results from marine (salt water) conditions are necessary to determine the
versatility of this device. The sample set is small due to lack of funding for this project.
Additional studies are needed to determine the repeatability and reliability of this device.
Additional pathogenic serotypes, as well as, total E. coli could be adapted to the
biosensor and studies need to be conducted to test such configurations. Finally, since this
study was conducted on a river, a similar study needs to be conducted for beaches located
on lakes. With additional studies and implementing the findings from these future
studies, in the near future the biosensor may provide the best available method for
determining the safety of recreational use water. Image this; a lifeguard dips a biosensor
type device into the lake water. This device tests the water for the levels of many
different strains of bacteria. Using the biosensor device is as easy as testing for pH. The
results are available in minutes and the lifeguard can determine if the water is safe before

the beach is opened to the public.

147

Appendix

Table A-1. Farm Lane Three Years of Data.
Table A-2. Yearly Mean of E. coli Concentration, Flow and Rainfall

Table A-3. Nutrients, E. coli, and Land-use Data.

148

Table A-1. Farm Lane Three Years of Data.

Week

@ONO’UI4DUN-l

Year

2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001
2001

E. coli

350

1 13
170
41 O
383
1273
207
670
360
333
1 333
490
493
393
560
690
270
393
357
287
1 580
693
5833
550
267
7600
93
90
85
683
170
170
720

thee 300 thes1000 Flow

A‘OOOO‘AA-Ao-rhoodoc-deAOA—l—hAAOOOO‘OOAOOO-AO—h—l—h-LO-fi—lo-b..h—t-l-h-A-h—h‘O‘AAOOA

OOOOOO-b-lOOOAOOOOO-IOOOOOOJd-bOOOOOOOOOOOAOOJOAOOOOOOOCOAOOOO-‘OOOOO

Rain 24

546
157
1 13
267
784
275
173
149
221
180
423
124

92

53
152

149

.09
do
ma

0
3.0
Muooooo

.0 .o .o .o .09
ooouoo¢o¢ooo-oo OOOOOOO’OQ COO

.0
01

99
an
‘0’

0

Rain 48

0.01

0.01

0.02

Rain 72

0.23

1.33

0.52

temp hi
61.6
69.4
82.5

temp low humd hi

35.6

64.7

humd lo

Table A-1. Farm Lane Three Years of Data. (continued)

Week

omwmmeuN-s

Year

2002
2002
2002

2002
2002
2002
2002
2002
2002
2002
2002
2002
2002
2002

F. coli

163

273

thee 300 thes 1000 Flow

O-‘OOOOOOOOJOOA-AO-‘AOO-AOOO-b-lo44000

OOOOOOOOOOOOOC-‘CO-‘OO-‘OOOO-‘OO-IOOO

Rain 24

150

0.1
0.21
0.14
0.01

0.01
0.78
0.01

.o .o
o uh
V

O
C
OONOOOOOUOOOOO

Rain 48 Rain 72

0.11
0.39
0.14
0.81

0.01

0.2
0.39
0.14
0.91
0.03
0.14
0.37

0
0.15
O

0
0.2
0
0.01
1.23
0.02
O
0.17

temp hi
44.7
55.9
85.4
51 .3
46.3
79.1
61.8
85.5
70.5
89.8
92.5
90.9
88.3
91.2
86.5
8108
88.5
70.8
82.5
82.5
90.5
75.9
65.2

temp low humd hi

32.8
36.1
56.4
42.4
31.4
43.8
46.6

55
52.1
65.6
67.9
59.4
55.1

humd 10
46.9
85.5
51.5
54.3
39.7
23.9
56.4
36.6
38.9
37.6
41.8
31.7
40.3
48.7
66.8
50.8
41.3

63
39.7

49
33.5
30.5
32.5
48.2
32.2
28.5
39.5
46.2
39.6
78.5
51.5
55.8

Table A-1. Farm Lane Three Years of Data. (continued)

soil moist 1
0.287
0.264
0.205
0.336
0.355
0.362
0.385
0.368
0.356
0.379
0.362
0.332
0.356
0. 286
0. 343
0.358
0.278
0.267
0.292
0.218
0.341
0.301
0.316
0.274
0.297
0.276
0.247
0.298
0.268
0.278
0.284
0.296
0.303
0.222
0.178
0.238
0.323
0.353
0.356

0.287
0.282
0.191
0.191

0.18
0.191
0.313
0.191

0.18
0.315
0.247

0.18
0.376
0.191
0.381
0.308
0.336
0.402
0.314
0.347
0.341
0.371

0.348

eoilmoe‘llo eoilmotetz eolmerlo eolternphl eoiltemplo eolerred widened dudrcomt
0.468

0.271 0.579 60.5 429 374.9 13.2
0.238 0.773 0.573 66.4 46.6 644.3 17.9
0.191 0.571 0.492 71.6 61 409.7 23.3
0.321 0.88 0.772 60.4 51.8 598.7 31.7
0.337 0.895 0.73 71.1 53.1 555.2 12.5
0.344 0.928 0.759 75 53.4 677.8 12.5
0.347 0.949 0.789 71.2 56.2 453.8 9.8
0.316 1.237 0.538 75.9 69.7 275.3 19.2
0.347 1.192 0.899 78.5 64.5 346.8 10.5
0.359 1.161 0.936 85.3 67.8 563 14.5
0.351 1.204 0.916 86.1 62.7 592.5 11.2
0.319 1.041 0.867 77.8 64.5 335.2 25.9

0.34 1.205 0.977 81.6 63.6 521.2 13.9
0.272 1.06 0.821 83.2 61 505.4 11.2
0.307 1.588 75 68.5 158.8 16.6

0.33 0.795 76.9 65.7 302.2 22.6
0.273 7.38 0.332 75.9 64.4 309.2 13.2
0.258 0.102 69.6 57.1 620.5 13.9
0.279 7.43 3.372 71.7 66.9 364.1 15.9
0.198 5&3 0.256 70.4 61.8 309.6 23.9
0.191 7.88 4.53 71.1 68.2 49.4 13.5
0.292 6866 0.056 67.8 54.2 462.4 15.9
0.307 2.7 1.838 60.7 54.5 364.6 16.6
0.269 2.292 1.341 64.3 53.9 386.5 22.6
0.289 3.403 2.127 47.4 43.9 225.5 18.6
0.258 2.226 1.331 60.6 52.4 164.3 12.5
0.239 2.276 1.275 57.9 49.5 282.5 13.9

0.29 0.471 0.458 52.2 43.1 304.7 12.2
0.261 0.454 0.42 50.9 41.8 247.6 10.2
0.275 0.469 0.468 45 43.2 38 12.9
0.282 0.469 0.466 36.4 34.8 135 22.3
0.286 0.494 0.468 42.5 37.8 41.7 11.2
0.265 0.491 0.418 62.2 50.6 302 20.2
0.207 0.386 0.34 62.7 45.3 572.4 15.9
0.168 0.275 0.25 63.9 53 490.6 20.6
0.229 0.375 0.324 69.5 47.2 606.1 16.9
0.322 7.23 0 70.4 60.2 189.8 30.3
0.341 7.63 0.155 81.3 54 488.6 19.6
0.343 7.46 0.6.32 72 53.3 256.8 19.2
0.282 6.963 0.498 ”.1 59.6 397.7 19.9

0.26 6.723 2.567 92.8 60.3 694 20.6
0.255 6.59 2.118 94.6 58.3 695.3 12.5
0.191 6.708 5.056 92.7 64.1 534.2 24.3
0.187 6.447 4.179 104 69.8 653.6 12.9
0.179 5.196 3.585 99.4 64.2 532.7 13.5
0.191 4.035 2.816 102.9 69.7 545.9 12.9
0.187 4.193 2.943 101.4 69.1 493.4 37.7
0.179 3.391 2.318 105.1 69.3 588.1 12.5
0.179 3.407 2.462 97.1 68.3 419 14.9
0.267 4.921 3.94 60.6 64.5 270.9 15.5
0.215 5.314 4.045 88.8 68.6 314.1 17.2
0.179 5.078 3.436 91 57. 6 524.8 13.9
0.277 6. 497 3.69 81.2 65. 6 261.8 25.3
0.191 5. 996 3.367 75 57.3 309.7 13.5
0.338 5.014 3.506 51.5 45.9 126.1 18.2
0.287 4&5 2.692 65.5 45.4 421.6 9.5
0.321 5.253 1.193 51 39 311.4 20.9
0.355 4.707 3.466 61.8 51 95.1 27.3
0.306 5.193 1.313 59.2 48.6 197.4 14.9
0.339 1.816 0.206 48.5 35.8 281.6 20.6
0.332 5.281 0.638 51.2 41.9 254.7 22.3
0.299 0. 785 0 47.9 34.6 224.4 12.9
0.329 0. 956 0.91 52.7 40.1 122.4 14. 2
0.343 0.957 0.933 48.4 43.1 118.9 16. 6

151

NOODOOOUOOO‘dOON-INOU

.5
.0000!

OOOOOOOO

water temp Week

14.07
13.76
15.97
14.23
16.44
13.92
12.66
16.04
20.63
19.33
19.01
21.63
21.29
25.55
21.79
24.16
22%
19.66
21.29
18.66
20.14

16.6
10.”
11.91

8.19
13.15
10.21

6.64

8.19

5.71
10.05

8.19

OONO‘fl‘UN-e

YOU

Table A-1. Farm Lane Three Years of Data. (continued)

soil moist1soil most eoilmoietZsoilmosz eoiltempheoiltempkeolarrad windepeecdudtcountwatertempWeek

0.243
0.31
0.281

0.227
0.273
0.238

0.257
0.256
0.322

0.267
0.215
0.284
0.301
0.354
0.328
0.295
0.322
0.367
0.333

0.256

3.11
0.286
0.314
0.297
0.286
0.309
0.275
0.331
0.294
0.304

0.804
0.758
0.781
0.902
0.859
0.753
0.695
0.654

1.26
1.206
1.183
1.167
1.506
1.558
1.673
1.529
1.477
1.582
1.677
1.595
1.536
1.425
1.386
1.301
1.468
1.434
1.408
1.505
1.477

1.74
1.747
1.769

0.761
0.696
0.745
0.865
0.821
0.703
0.606
0.621
1.224
1.156
1.137
1.126
1.444
1.485
1.505
1.466
1.425
1.502
1.609

1.55
1.486
1.356
1.331
1.314
1.427
1.386
1.362
1.471
1.432
1.698
1.696

1.75

54.7
51.8
61.4
54.3
49.9
66.1
62.2
71.2
64.5
78.8

43.7
43.3
54.2
49.6
43.1
50.8
56.9
59.2
67.9
67.8
71.3
69.9
65.6
69.1
72.3
70.2
70.7
64.6
64.9
69.3
66.7
57.4
53.1
59.3
49.5
43.3
39.5
41.1
37.2
45.1
32.9
36.5

152

20.6
30
22.3
17.2
13.5
14.5
22.3
10.5
21.9
15.9
16.9
26.3
16.6
19.2
20.9
18.9
18.6
14.2 x
10.5
19.2
12.9
12.5
15.2
15.9
24.3
13.9
14.5
15.2
15.2
14.9
17.6
15.5

8.34

OGNOG¥~UM4

Year

 

Table A-2. Yearly Mean of E. coli Concentration, Flow and Rainfall

 

 

 

 

 

 

 

 

 

 

 

Year E. coli Flow Rain 24h Rain 48h Rain 72h
'2'000 Mean 854.656 207.437 .1328 .2528 .4300
N 32 32 32 32 32
Std. 1594.15 175.2794 .3194 .4131 .5802
2001 Mean 872.796 208.312 .0950 .1684 .3412
N 32 32 32 32 32
Std. 1226.14 180.0730 .1754 .2954 .4322
2002 Mean 530.406 51.140 .0641 .1347 .1556
N 32 32 32 32 32
Std. 894.4360 56.8654 .1563 .2599 .2766
Total Mean 752.619 155.630 .0973 .1853 .3090
N 96 96 96 96 96
Std. 1267.25 164.8587 282 .3297 .4571

 

153

 

Table A-3. Nutrients, E. coli, and Land-use Data.

Site Week

SSSSSSSSSSSSS

Date

4125/2001
5/9/2001
512312001
61612001
612012001
711 812001
811/2001
811512001
8/29/2001
912512001
101812001
1 11612001
1112012001
412512001
51912001
512312001
61612001
612012001
711812001
811/2001
811512001
812912001
912512001
101812001
ӣ12001
1112012001
412512001
519/2001
5123/2001
6/6/2001
612012001
711812001
81112001
811512001

§§§§§§§8§3§S§§§§§§§§

N
(I
.5
O

§§§§§§§§

85

TP

0.855
0.091
0.013
0.023
0.019
0.001
0.009
0.091
0.018

154

Rain 24

0
0.04
0
0.02
0
0.04
0
0
0.06
0
0.14
0
0.13
0
0

P
c
we

9 .o
“ 8
GOOOOOO OOOOO

.°
9

. .° .°
" 8
accoooo ooocc

O
O

.o .o
" 8
GOOOOOO OOOOO

.0
o

.o .o
" 8
000 000°C

Rain48 Rain72

0.31
0.72
0.44
0.02
0.18
0.04

0

0
0.58

0
0.14

0
0.13
0.08
0.45
0.35
0.03
0.12

0.74 Aar
0.72 Agr
0.44 Aar
0.02 Apr
0.18 Aor
0.04 Apr
0.38 Apr
0 Apr
0.58 Aor
0.62 Aar
0.14 Apr
0 M!
0.13 Mr
0.34 Mixed
0.45 Mixed
0.66 Mixed
0.07 Mixed
0.13 Mixed
0 Mixed
1.03 Mixed
0 Mixed
0.1 Mixed
0.37 Mixed
0.62 Mixed
0 Mixed
0.19 Mixed
0.34 Aor
0.45 Aar
0.66 Aor
0.07 Apr
0.13 Apr
0 Mr
1.03 Act
0 Act
0.1 Aor
0.37 Aor
0.62 Act
0 M
0.19 Act
0.34 Mixed
0.45 Mixed
0.66 Mixed
0.07 Mixed
0.13 Mixed
0 Mixed
1.03 Mixed
0 Mixed
0.1 Mixed
0.37 Mixed
0.62 Mixed

Land use

Table A-3. Nutrients, E. coli, and Land-use Data. (continued)

Site Week Date E. coli Rain 24 Rain 48 Rain 72 Land use
81 1 4125/2001 0 0.51 0.98 urban
S1 3 51912001 1 0.04 0.54 0.54 urban
s1 5 5123/2001 0 0.05 0.05 urban
s1 7 61612001 0.06 0.06 0.06 urban
s1 9 6120/2001 0 0.76 1 urban
S1 13 7118/2001 0 0 0 urban
S1 15 8115/2001 0 0 0 urban
$1 17 8129/2001 0 0.03 0.03 urban
S1 19 9125/2001 0.03 0.71 0.71 urban
S1 23 1018/2001 0 0 0.2 urban
S1 25 1012312001 0.31 0.31 0.31 urban
$1 29 1116/2001 0 0 0 urban
S1 31 1112012001 0.14 0.14 0.15 urban
S19 1 4125/2001 0 0.51 0.98 Mixed
S18 3 51912001 0.04 0.54 0.54 Mixed
S18 5 5123/2001 0 0.05 0.05 Mixed
S18 7 61612001 0.06 0.06 0.06 Mixed
S18 9 6120/2001 2237 0 0.78 1 Mixed
S18 13 7118/2001 0 0 0 Mixed
$18 15 81112001 0 0 0 Mixed
$18 17 8115/2001 0 0.03 0.03 Mixed
S18 19 9125/2001 0.03 0.71 0.71 Mixed
$18 23 101812001 0 0 0.2 Mixed
S18 25 1012312001 0.31 0.31 0.31 Mixed
s18 29 1116/2001 0 0 0 Mixed
$18 31 1112012001 0.14 0.14 0.15 Mixed
$28 1 4125/2001 0 0.51 0.96 urban
$28 3 51912001 0.04 0.54 0.54 urban
$28 5 5123/2001 0 0.05 0.05 urban
$28 7 61812001 0.06 0.08 0.06 urban
$28 9 612012001 1 0 0.78 1 urban
$28 13 7118/2001 0 0 0 urban
$28 15 611/2001 11 0 0 0 urban
$28 17 8115/2001 0 0.03 0.03 urban
$28 19 9125/2001 0.03 0.71 0.71 urban
828 23 1018/2001 0 0 0.2 urban
$28 25 1012312001 0.31 0.31 0.31 urban
S28 29 1118/2001 0 0 0 urban
$28 31 11/20/2001 0.14 0.14 0.15 urban
$5 1 4125/2001 0 0.31 0.74 Mr
85 3 51912001 0.04 0.72 0.72 Apr
85 5 5123/2001 0 0.44 0.44 Apr
85 7 81612001 0.02 0.02 0.02 Apr
85 9 6120/2001 0 0.18 0.18 Agr
$5 13 7118/2001 0.04 0.04 0.04 Apr
85 15 81112001 0 0 0.38 Apr
85 17 8115/2001 0 0 0 Agr
$5 19 9125/2001 0.06 0.58 0.58 Apr
85 23 1018/2001 0 0 0.62 Apr
85 25 1012312001 0.14 0.14 0.14 Apr
85 29 1116/2001 0 0 0 Apr
85 31 1112012001 1 0.13 0.13 0.13 Apr
88 1 4125/2001 0 0.31 0.74 Apr
88 3 51912001 0.04 0.72 0.72 Apr
88 5 5123/2001 0 0.44 0.44 Apr
88 7 61812001 0.02 0.02 0.02 Apr
88 9 6120/2001 1 0 0.18 0.18 Agr
$8 13 711812001 0.04 0.04 0.04 Apr
88 15 81112001 0 0 0.36 Agr
S8 17 8110/2001 0 0 0 Apr
88 19 9125/2001 0.06 0.58 0.58 Apr
88 23 1018/2001 0 0 0.82 Apr
88 25 1012312001 0.14 0.14 0.14 Apr
88 29 1118/2001 0 0 0 Apr
88 31 1112012001 0.13 0.13 0.13 Agr

 

155

Table A-3. Nutrients, E. coli, and Land-use Data. (continued)

Site
810
810
510
810
$10
$10
810
$10
$10
810
$10
810
$10
$11
$11
$11
$11
$11
$11
$11
$11
811
S11
811
$11
811
S13
$13
813
$13
$13
813
$13
$13
$13
$13
813
$13
813
$14
$14
814
S14
$14
814
814
$14
$14
$14
$14
814
$14
$16
$16
$16
$16
$16
816
S16
$16
$16
$16
$16
816
$16
817
S17
S17
S17
S17
$17
$17
$17

Week

23
25
29
31
1
3
5
7
9
13
15
17
19
23
25
29
31
1

Date

412512001
51912001
512 312001
61612001
612012001
711 812001
81112001
811 512001
9125/2001
1 01812001

1 012312001
1 11612001

1 1 12012001
412512001
51912001
512 312001
61612001
612012001
711 812001
61112001
811 512001
912 512001

1 01812001

1 012 312001
1 1 1612001

1 112012001
412 512001
51912001
512312001
61612001
612012001
711 812001
811 12001
811 512001
912 512001

1 01812001

1 012 312001
1 11612001

1 1 12012001
412 512001
51912001
512 312001
61612001
612012001
711812001
61112001
811 512001
912512001

1 01812001

1 012 312001
1 11612001

1 1 12012001
412 512001
51912001
512 312001
61612001
612012001
711 812001
81112001
611 512001
9125/2001

1 01612001

1 012312001
1 11612001

1 1 12012001
412 512001
51912001
512 312001
61612001
612012001
711 812001
811 12001
611 512001

E. coli

3833

10533

2643

1007

2500
2000
1800
1700

4900

 

0.093
0.1
0.003
0.035
0.12
0.005
0.1 1
0.041
0.014
0.026
0.023
0.046
0.039
0.033
0.05
0.003
0.018
0.031
0.005
0.096
0.008
0.006
0.006
0.026
0.028
0.014
0.039
0.014
0.001
0.023
0.05
0.005
0.125
0.014
0.009
0.02
0.056
0.03
0.019
0.044
0.026
0.001
0.035
0.056
0.005
0.101
0.01 1
0.015
0.014
0.028
0.033
0.013
0.1 1
0.086
0.031
0.068
0.076
0.001
0.106
0.026
0.363
0.036

0.025
0.05
0.085
0.091
0.008
0.014
0.1
0.021
0.331
0.491

156

 

Rain 24
0
0.04
0
0.02
0
0.04
0
0
0.06
0
0.14
0
0.13
0
0.04
0
0.02
0
0.04
0
0
0.06
O
0.14
0
0.13
0
0.04
0
0.02
0
0.04
0
0
0.06
0
0.14
0
0.13
0
0.04
0
0.02
0
0.04
0
0
0.06
0
0.14
0
0.13
0
0.04
0
0.02
0
0.04
0
0
0.06
0
0.14
0
0.13
0
0.04
0
0.02
0
0.04
0
0

Rain 46
0.31
0.72
0.44
0.02
0.16
0.04

0
0
0.56
0
0.14
0
0.13
0.31
0.72
0.44
0.02
0.16
0.04
0
0
0.58
0
0.14
0
0.13
0.31
0.72
0.44

Rain 72

0.74 Mixed
0.72 Mixed
0.44 Mixed
0.02 Mixed
0.18 Mixed
0.04 Mixed
0.38 Mixed

0 Mixed
0.56 Mixed
0.62 Mixed
0.14 Mixed

0 Mixed
0.13 Mixed
0.74 urban
0.72 urban
0.44 urban
0.02 urban
0.18 urban
0.04 urban
0.38 urban

0 urban
0.56 urban
0.62 urban
0.14 urban

0 urban
0.13 urban
0.74 Agr
0.72 Apr
0.44 Agr
0.02 Agr
0.16 Apr
0.04 Agr
0.38 Apr

0 Apr
0.58 Agr
0.62 Agr
0.14 Apr

0 Apr
0.13 Apr
0.74 Apr
0.72 Agr
0.44 Agr
0.02 Agr
0.18 Agr
0.04 Agr
0.38 Apr

0 Apr
0.58 Apr
0.82 Agr
0.14 Apr

0 Apr
0.13 Apr
0.74 Apr
0.72 Agr
0.44 Apr
0.02 Apr
0.18 Apr
0.04 Agr
0.38 Apr

0 Apr
0.58 Apr
0.62 Agr
0.14 Apr

0 Apr
0.13 Agr
0.74 Mixed
0.72 Mixed
0.44 Mixed
0.02 Mixed
0.16 Mixed
0.04 Mixed
0.38 Mixed

0 Mixed

Land use

  
   

16AM STATE UNIVERSITY laflARlES

11111111111111) 111111111

02504 001