United States
Environmental Protection
Agency
Research and Development
Municipal Environmental Research EPA-600, 2 79 206
Laboratory December 1 979
Cincinnati OH 45268
Water Filtration for
Asbestos Fiber
Removal
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-206
December 1979
WATER FILTRATION FOR ASBESTOS FIBER REMOVAL
by
Gary S. Logsdon
Drinking Water Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention, treat-
ment, and management of wastewater and solid and hazardous waste pollutant
discharges from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and for minimizing the adverse
economic, social, health, and aesthetic effects of pollution. This publi-
cation is one of the products of that research; a most vital communications
link between the researcher and the user community.
This report shows that granular media filtration and diatomaceous earth
filtration can remove as much as 99.99 percent of the asbestos fibers from
drinking water. Fiber counts below the limit of detection by the transmis-
sion electron microscope method can often be achieved. The report presents
results of research conducted at Duluth, Minnesota and Seattle, Washington
as well as monitoring data from Silver Bay and Two Harbors, Minnesota;
Philadelphia, Pennsylvania; Chicago, Illinois; and the San Francisco Bay
area in California. Even though data were collected under diverse condi-
tions, the relationship of low turbidity and low fiber count in filtered
water is seen as a unifying factor.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
iii
-------
PREFACE
The Safe Drinking Water Act (P.L. 93-523) gives the Administrator of
the Environmental Protection Agency authority to issue regulations if a
substance may pose a health hazard in drinking water. Inhaled asbestos
is a known carcinogen, and ingested asbestos may be hazardous to health.
Therefore information about asbestos in drinking water should be of interest
to public health officials, the water utility industry, and the Environmental
Protection Agency.
During the 1970's the results of electron microscope examination of
water samples showed that some waters contained asbestos fibers. Early
Canadian efforts created both concern and curiosity in other investigators,
and gradually knowledge about asbestos fibers in drinking water increased.
Because the fibers might constitute a health hazard in drinking water,
monitoring efforts were carried out to assess the capability of water fil-
tration plants to remove asbestos, and engineering research was funded to
provide more detailed knowledge about water filtration for asbestos
removal.
Information about asbestos in drinking water and about fiber removal
by filtration is scattered throughout the engineering and scientific liter-
ature. Results have been reported by numerous investigators, concerning a
variety of waters and locations, based upon a number of different variations
of analytical technique. Because of such diversity, those who review the
literature can expect to find results that appear contradictory or difficult
to explain. Therefore the Drinking Water Research Division has undertaken a
comprehensive review of asbestos in drinking water, in order to summarize
and interpret the work of the many investigators who have made contributions
to present day knowledge of the problem.
The goal of this report is to present a unified concept of water filtra-
tion for asbestos fiber removal that can be used to explain past results as
well as to guide filtration plant designers as they plan new facilities and
to help water filtration plant operators as they strive to reduce the concen-
tration of asbestos fibers in drinking water.
IV
-------
ABSTRACT
This report was prepared so that concepts of water filtration for asbestos
fiber removal could be presented in a manner that would explain results
reported by other investigators, to provide information for designers of water
filtration plants, and to give guidelines for the successful operation of
filtration plants that are removing asbestos fibers from drinking water. This
document reviews the literature of other investigators, (mostly Canadian),
summarizes filtration studies funded by U.S. EPA at Duluth and Seattle, and
presents monitoring data gathered at water treatment plants in Philadelphia,
Chicago, and the San Francisco Bay area.
Results of treatment for asbestos removal presented herein cover not only
a number of widely separated geographic locations from coast to coast in the
United States and Canada, but also from a wide range of source waters, from a
pristine mountain lake to turbid rivers and estuarine waters, as well as
variations in flow from 10 gallons per minute or smaller to more than -200
million gallons per day and several modifications of granular media filtra-
tion. Although the results are presented on the basis of geographic location,
the discussion of results is based upon the type of filtration used, because
of the substantial differences between granular media filtration and diato-
maceous earth filtration.
Data from Seattle and Duluth show that chrysotile and amphibole fiber
concentrations in drinking water can be substantially reduced by granular
media filtration. Reductions of up to 99.99 percent were reported during
storm conditions at Duluth, Minnesota. Effective granular media filtration
required very diligent plant operation with careful control of pH, coagulant
doses, and filtered water turbidity.
Even though turbidity can not directly measure asbestos fibers in the
concentrations found at water treatment plants, when a granular media filtra-
tion plant is properly operated, turbidity readings can be used as a guide to
plant operation. Filtered water turbidity should be 0.10 nephelometric
turbidity units (ntu) or lower to maximize fiber removal. Turbidity increases
of 0.1 or 0.2 ntu above this value generally were accompanied by large
increases in asbestos fiber concentrations.
Some of the results presented suggest that fiber removal is more easily
accomplished when source waters have turbidities greater than 1 ntu, the raw
water turbidity typical at Duluth and Seattle.
Diatomaceous earth filtration was found effective for asbestos fiber
removal in bench-scale and pilot plant studies. A full scale diatomite fil-
tration plant has not yet been evaluated for fiber removal efficiency.
-------
Research to date indicates that coating the diatomaceous earth filter aid with
aluminum hydroxide substantially increases the removal of both amphibole and
chrysotile fibers. Duluth results indicate that filtered water turbidity
should be 0.10 ntu for most effective fiber removal.
The principal studies on which this report is based were performed during
the period from March, 1974 to June, 1979. This report was completed in
September 1979.
vi
-------
CONTENTS
Foreword ill
Preface iv
Abstract v
Figures viii
Tables x
List of Metric Conversions xiii
Acknowledgments xiv
1. Introduction 1
2. Conclusions 2
3. Recommendations 4
4. Review of Literature 5
5. Water Utilities on Lake Superior's North Shore 12
6. Seattle Pilot Plant Study 33
7. San Francisco Bay Area 59
8. Philadelphia 74
9. Chicago 84
10. Discussion 87
References 126
Appendices
A-l. Duluth Monitoring Data from Lakewood Filtration
Plant 131
A-2. Effects of Filtration Rate Changes on Water
Quality at Duluth 138
B. Summary of Seattle Pilot Plant Asbestos Data 141
vii
-------
FIGURES
Number Page
1 Flow Diagram for Filtration Plant at Duluth, Showing
Treatment Train Options 20
2 Lakewood Filtration Plant Turbidity Data , 22
3 Lakewood Filtration Plant Amphibole Fiber Data. . . . 24
4 Flow Diagram for Filtration Plant at Two Harbors 28
5 Flow Diagram for Filtration Plant at Silver Bay 31
6 Pilot Plant Flow Schematic - Seattle 36
7 Flow Diagram for Filter Columns - Seattle Pilot Plant 38
8 Finished Water Turbidity and Aluminum Residual vs. pH -
Seattle Pilot Plant 45
9 Finished Water Turbidity vs. Alum Dosage - Seattle Pilot Plant. . . 46
10 Finished Water Turbidity vs. Catfloc Dosage -
Seattle Pilot Plant , . . f 49
11 Finished Water Turbidity vs. 573C Dosage - Seattle
Pilot Plant 50
12 Net Water Produced per 24 House vs. Filter Loading Rate 56
13 Water Sources for Treatment Plants in Contra Costa County 60
14 Water Sources for Marin Municipal Water District Plants 70
15 Flow and Turbidity During Torresdale Plant Storm Sampling 78
16 Flow and Turbidity During Belmont Plant Storm Sampling 79
17 Chicago Asbestos Monitoring Data - Jardine Water
Filtration Plant 85
18 Chicago Asbestos Monitoring Data - South Water Filtration Plant . . 86
viii
-------
19 Raw Water Chrysotile vs. Turbidity - Seattle Pilot Plant 89
20 Raw Water Amphibole vs. Turbidity - Seattle Pilot Plant 90
21 Relationship between Raw Water Turbidity at Duluth Lakewood
Intake and ORF Amphibole Fiber Counts 92
22 Finished Water Turbidity and Chrysotile vs. Time,
Run #21 - Seattle Pilot Plant 93
23 Finished Water Turbidity and Chrysotile vs. Time,
Run #24 - Seattle Pilot Plant 94
24 Operating Data for Run # 174MM - Seattle Pilot Plant 95
25 Finished Water Turbidity and Chrysotile vs. Time,
Run #120 - Seattle Pilot Plant 96
26 Effluent Turbidity vs. Amphibole Fiber Count,
Duluth Granular Media Pilot Plants' 100
27 Amphibole Count vs. Finished Water Turbidity -
Seattle Pilot Plant 102
28 Chrysotile Count vs. Finished Water Turbidity -
Seattle Pilot Plant 103
29 Comparison of Frequency Distributions of Fiber Counts of
Filtered Water Sample Groups, Lakewood Plant - Duluth 104
30 Frequency Distributions of Amphibole Fiber Counts for Filtered
Waters from Plants on Lake Superior 105
31 Frequency Distribution of Chrysotile Fiber Counts for Filtered
Water Sample Groups - Seattle Pilot Plant 106
32 Relationship of Turbidity and Chrysotile Removal by Granular
Media Filtration - Seattle Pilot Plant 109
33 Effluent Turbidity vs. Amphibole Fiber Count for Pressure
Diatomite Filtration - Duluth Pilot Plant 118
34 Direct Filtration Plant Capacity vs. Cost 121
35 Operation & Maintenance Cost vs. Water Production for
Direct Filtration (May 1979) 123
ix
-------
TABLES
Number Fa§e
1 Asbestos Count Correlations - Everett Pilot Plant ........ 10
2 Amphibole Mass and Fiber Removal by Dual Media Filtration -
Duluth Pilot Plant ...................... 14
3 Amphibole Mass and Fiber Removal by Mixed Media Filtration -
Duluth Pilot Plant ...................... 15
4 Amphibole Mass and Fiber Removal by Diatomite Filtration -
Duluth Pilot Plant ...................... 16
5 Comparison of Raw and Filtered Water Chrysotile Data for
Data Pairs with Filtered Water Turbidity Equal to or
Less than 0.10 ntu - Duluth Pilot Plant ........... 18
6 Plant Hydraulic Information - Duluth .............. 19
7 Duluth Filter Media ....................... 21
8 Duluth Filtration Plant Performance During Storms ........ 25
9 Amphibole Fiber Concentration in Duluth Distribution
System Reservoirs ...................... 26
10 Plant Hydraulic Information - Two Harbors ............ 29
11 Two Harbors Amphibole Data ................... 30
12 Plant Hydraulic Information - Silver Bay ............ 32
13 Silver Bay Amphibole Data .................... 32
14 Dimensions of Unit Processes - Seattle Pilot Plant ....... 37
15 Velocity Gradient Data - Seattle Pilot Plant .......... 39
16 Mixing Intensities for Flocculator - Seattle Pilot Plant .... 49
17 Characteristics of Media Tested - Seattle Pilot Plant ...... 41
18 Raw Water Quality Characteristics - Seattle ........... 42
-------
19 Raw Water Asbestos Counts - Seattle Pilot Plant 43
20 Seattle Pilot Plant Preferred Chemical Treatments - Alum .... 44
21 Alum and Lime Dosage vs. Cationic Polymer Dosage -
Seattle Pilot Plant 47
22 Seattle Pilot Plant Preferred Chemical Treatments -
Alum Plus Cationic Polymer 47
23 Catfloc T-l Dosages and Turbidity - Seattle Pilot Plant 48
24 573 C Dosages and Turbidity - Seattle Pilot Plant 51
25 Comparison of One vs. Three Static Mixers -
Seattle Pilot Plant 52
26 Results from Filter Runs with and without Flocculation -
Seattle Pilot Plant 54
27 Chrysotile Results from Various Filter Media -
Seattle Pilot Plant 55
28 Summary of Fiber Count Data for Seattle Study 58
29 Design and Hydraulic Information for Bollman Plant,
Contra Costa County Water District 62
30 Bollman Plant Filter Media 62
31 Raw Water Data for Contra Costa Canal and Other Surface
Waters in Contra Costa County 63
32 Treatment Data for Antioch Plant 65
33 Treatment Data for Bollman Plant 66
34 Treatment Data for Pittsburg Plant 68
35 Treatment Data for San Andreas Plant 69
36 Treatment Data for Marin Municipal Water District
and North Marin Water District 71
37 Design and Operating Information for
Philadelphia Plants, 1977-1978 76
38 Storm Flow Sampling Results at Philadelphia 80
39 Water Quality Data - Torresdale Plant 81
xi
-------
40 Water Quality Data - Queen Lane Plant 82
41 Water Quality Data - Belmont Plant 83
42 Seattle Pilot Plant Chrysotile Fiber Counts In
Filtered Water When Turbidity Rose Above 0.10 ntu 98
43 Release of Asbestos Fibers by Filters at
Seattle Pilot Plant 98
44 Chi Square Analysis of Fiber Removal and Turbidity -
Seattle 108
45 Treatment Plant Capital Costs 120
46 Water Filtration Plant Operation & Maintenance Costs 122
47 Chemical Doses for Asbestos Fiber Removal 125
xii
-------
LIST OF METRIC CONVERSIONS
Filter area
Filter box depth
Filter head loss
Filtration rate
Flocculator paddle velocity
Plant capacity
River flow rate
Torque
Unit costs
Volume
Water production of a filter
Water flow in treatment process
Weir overflow
1.0 ft2 = 0.093 m2
1.0 ft = 0.30 m
1.0 ft of water = 0.30 m water
1.0 gal/min/sq ft = 2.4 m/hr
1.0 ft/sec = 0.30 m/sec
1.0 mgd = 3800 m3/day
1.00 cu ft/sec = 0.028 m3/sec
1.0 ft-lb = 1.36 N-m
1.0<1000 gal = 0.26<
-------
ACKNOWLEDGMENTS
This report is based upon the efforts of many persons and a large number
of organizations, but special consideration is due to the Canadian scientists
and engineers who conducted the first studies that revealed the problem of
asbestos in drinking water and suggested that water filtration could provide
the solution.
Contributors to the effort at Duluth and the Lake Superior north shore
include Black & Veatch, the Ontario Research Foundation, EPA Environmental
Research Laboratory at Duluth, University of Minnesota-Duluth, Lake Superior
Basin Studies Group, RREM Inc. Consulting Engineers, and the Duluth Depart-
ment of Water and Gas.
The Seattle pilot plant study was performed by the Seattle Water Depart-
ment. Electron microscope analysis of water samples was done by the Univer-
sity of Washington.
Data for the San Francisco Bay area came from research done at the
University of California, Berkeley, with much supplemental treatment plant
data furnished by the Sanitary Engineering Section, Department of Health,
State of California. Asbestos samples were analyzed at the University of
California, Berkeley.
Chicago data were furnished by the Chicago Department of Water and
Sewers. The Chicago Water Purification Laboratory analyzed the samples for
asbestos.
The Philadelphia Water Department provided information on Philadelphia's
water treatment plants. Asbestos samples from Philadelphia were analyzed at
the University of Washington.
xiv
-------
SECTION 1
INTRODUCTION
Asbestos is used in many ways in the world today. It has found many
successful commercial and industrial applications for numerous decades. Over
a period of years, public health officials and medical experts have learned
that inhaled asbestos is an occupational health hazard. When Cunningham and
Pontefract (1) reported in 1971 that they had found asbestos fibers in drinking
water in Canada, asbestos came to the attention of regulatory officials and
the water utility industry. Health effects data for ingested asbestos fibers
were not available, so no basis for immediate regulatory action existed.
Because this was the first report of asbestos in drinking water, additional
investigation would be needed to describe the nature of the problem.
Asbestos in drinking water suddenly became a serious environmental
problem in the United States in the summer of 1973, as a result of studies
done at the Environmental Research Laboratory in Duluth (2). Involved in liti-
gation with the Reserve Mining Company, the Environmental Protection Agency
contended that amphibole fibers originating in tailings being dumped into Lake
Superior at Silver Bay, Minnesota were traveling with the lake current and
could be detected in the unfiltered drinking water of Duluth.
The Lake Superior problem stimulated much interest in asbestos in drinking
water, especially in the USA. Studies were made to improve analytical methods
for asbestos, to learn more about the health effects of ingested asbestos, and
to develop information about ways to remove asbestos fibers from drinking water
by filtration. Since 1973, much money and many hours of work have been ex-
pended to find some answers to the problem of asbestos in drinking water.
The purpose of this report is to assess the efficacy of water filtration
processes for asbestos fiber removal. This has been done by reviewing pub-
lished reports of work done by others and by studying in depth the results of
work done by or sponsored in some way by EPA. A heavy reliance is placed on
EPA-related work because generally these studies are more recent, and a greater
amount of additional engineering and water quality data could be obtained.
Having the engineering and water quality data is especially important for the
purpose of advising design engineers and water utility operators how to design
and run water filtration plants so that asbestos fibers can be efficiently
removed.
-------
SECTION 2
CONCLUSIONS
MONITORING
Analysis of water samples by transmission electron microscope is the only
method that provides asbestos fiber counts, but this technique can not be used
for control of treatment plant operation because of the long time delay be-
tween collection of samples and receipt of results.
A laser-illuminated optical detector has been developed and regression
equations can be used to convert detector signal output to estimates of
amphibole fiber count. This device could be adapted to on-line operation in
water filtration plants.
Turbidimeters do not detect asbestiform fibers in the range of concen-
trations of interest for drinking water - 0.01 to 1000 X 106 fibers/liter.
When water is properly conditioned for granular media filtration, changes of
turbidity of the filtered water usually signal changes in the fiber concen-
tration of filtered water.
GRANULAR MEDIA FILTRATION
i Asbestiform fiber removal by granular media filtration has exceeded 99.99
percent.
Proper conditioning of raw water is essential for effective fiber removal.
Adequate coagulant must be used to completely destabilize all particles.
Filter aids may be needed to prevent turbidity breakthrough.
After filtered water turbidity has stabilized at the recommended goal of
0.10 nephelometric turbidity-unit (ntu) or lower, an increase of even 0.1 ntu
in filtered turbidity is usually associated with a very large increase.in
filtered water fiber concentration, because granular media filters can store
and then slough large numbers of fibers in filter upset situations. Loss of
chemical feed and rapid filtration rate increases must be avoided. Careful
control of coagulation pH is needed, especially for low alkalinity waters.
Very careful attention to coagulation chemistry and filter operation will be
required for effective filtration plant operation..
Granular media filters can remove chrysotile fibers at filtration rates
of 2 to 10 gpm/sf. Depending on local circumstances,-both back mix and in-
line or static mix techniques can be effective. Some waters require floccu-
-------
lation for effective filtration. Sedimentation removes asbestos fibers but it
is not needed for relatively clear, 1 ntu or less, surface waters.
The principal types of granular media currently in use - sand,dual media,
and mixed media - have been shown to be effective for filtering out asbestos
fibers.
Designing filters with seven or more feet of water over the filter media
prevents air binding problems.
DIATOMACEOUS EARTH FILTRATION
Diatomaceous earth filtration can remove both amphibole and chrysotile
fibers, with demonstrated removals as high as 99.99 percent.
Diatomaceous earth (DE) coated with aluminum hydroxide is more effective
for removal of asbestos fibers than uncoated DE.
Filtered water turbidity should be 0.10 ntu or lower to attain maximum
filter efficacy for fiber removal.
Use of vacuum DE filters to treat cold, oxygen-saturated water may result
in formation of air bubbles in the filter cake. This problem can cause deter-
ioration of filtered water quality. Because of design differences, pressure
DE filters would notexperience the air bubble problem.
COSTS
Cost of filtration for surface waters containing asbestos fibers should
be very similar to the cost of filtering other surface waters. Meeting the
0.10 ntu goal for water quality is economically and technically feasible.
This is demonstrated by the ability of plants such as the Contra Costa County
Water District's Bollman plant and Duluth's Lakewood plant to meet such a
turbidity goal.
-------
SECTION 3
RECOMMENDATIONS
Water treatment plants that filter water for the purpose of removing
asbestos fibers should continuously monitor the effluent turbidity from each
filter.
Filtered water turbidity should be 0.10 ntu or lower.
A small increase in turbidity (0.1 to 0.2 ntu) may be associated with a
very large increase in asbestos fiber count in filtered water. For this
reason plant operators should backwash filters long before the 1 ntu Maximum
Contaminant Limit for turbidity is reached during turbidity breakthrough.
At granular media filtration plants, very careful control of chemical
feed rates should be maintained to assure that proper coagulant dose and pH
are maintained at all times.
Filtration rate should not be changed abruptly during a filter run.
Filters should be designed to avoid problems caused by formation of air
bubbles in the filter media. Use of pressure diatomaceous earth filters or
deep granular media filter boxes (about seven feet of water over media) is
recommended to prevent air binding problems in locations where they could
occur.
Usually a pilot plant study should be carried out before a water filtra-
tion plant is designed for asbestos fiber removal to help designers select
optimum filtration rates, coagulant dose and type, filter aid and dose and
type if needed, media type and so forth.
-------
SECTION 4
REVIEW OF LITERATURE
HEALTH EFFECTS
A considerable amount of controversy in the Reserve Mining Company trial
was related to the health effects of ingested asbestos. Asbestos is acknowl-
edged to be a carcinogen when inhaled, but in the early 1970's very little was
known about the health effects of ingested asbestos. Presenting a detailed
review of the health effects questions is not the purpose of this paper, but
a brief review of the health effects controversy will show why work on asbestos
in water is continuing. A more detailed paper on health effects was recently
prepared by McCabe and Millette (3).
One of the questions debated by scientists has been whether ingested
asbestos fibers could migrate through the gastrointestinal wall to the peri-
toneum and cause cancer. This possibility had been suggested because of the
association of peritoneal mesotheliomas with asbestos exposure, according to
Gross et. al. (4). The authors further stated that their report on ingested
mineral fibers was being published in response to the desire of public health
officials to learn whether or not ingested asbestos might present a health
hazard. In three separate laboratories investigators fed rats asbestos fibers
in butter or margarine, or introduced amosite fibers and taconite tailings
into the stomach of rats by means of a stomach tube.
Gross and the co-authors reported no incontrovertible proof of trans-
migration of the fibers. They also attacked a study by Pontefract and
Cunningham (5) in which the latter authors found amosite fibers widely dis-
seminated in tissues outside the gastrointestinal tract. Pontefract and
Cunningham had injected asbestos fibers through the stomach wall, and Gross
et. al. contended that the hole made by the needle allowed fibers to leak
from the stomach and contaminate the abdominal cavity. Thus controversy sur-
rounded the issue of the fate of ingested asbestos in the early 1970's.
In an attempt to resolve some of the controversy, EPA researchers
recently have sponsored or conducted studies. In a paper entitled "Ingested
Mineral Fibers: Elimination in Human Urine," Cook and Olson describe research
on the excretion of amphibole fibers in urine of persons drinking unfiltered
water (6). Cook and Olson found from 300 to 1200 amphibole fibers per milli-
liter (0.3 to 1.2 X 10^ fibers/liter) in the urine of persons who drank
unfiltered Lake Superior water. They estimated that the amphibole fiber
content of the unfiltered lake water ingested by the persons who voluntarily
provided urine samples would have been 50 X 10" (fibers/liter), or more in
certain cases. Cook and Olson stated, "These observations provide the first
-------
direct evidence for the passage of mineral fibers through the human gastro-
intestinal mucosa under normal conditions of the alimentary canal."
At the University of Illinois School of Public Health in Chicago,
researchers studied the fate of ingested chrysotile by bottle feeding a
suspension of chrysotile in milk formula to a newborn baboon that was kept
in an infant incubator under conditions of controlled temperature, humidity,
and air supply (7). The asbestos dose was about 3 X 10 ^ fibers per kilo-
gram of body weight. Chrysotile fibers were found in the kidney cortex
tissue of the exposed baboon, but were below limits of detection in the
unexposed control. Analysis of fiber size data showed very little difference
in the size frequency distributions for the fibers used in the study and the
fibers found in the cortex tissue. The investigators stated that no fiber
selection process seemed to be occurring ilti the gastrointestinal tract of
the newborn baboon. The maximum length of fiber found in the cortex tissue
was 35 urn.
The recent findings on the migration of asbestos fibers across the
gastrointestinal barrier do not in any way prove that the fibers cause gastro-
intestinal cancer or mesothelioma. The findings do, however, disprove the
argument that ingested asbestos could not cause peritoneal mesothelioma
because asbestos fibers do not migrate across the intestinal wall.
Because of findings like these, public health officials continue to be
concerned about the possible health hazards related to ingested asbestos.
The information on removal of asbestos fibers by water filtration processes
is presented in this context.
SOURCES OF ASBESTOS FIBERS IN DRINKING WATER
Asbestos fibers are contaminants in some drinking waters. They occur
when fibers contaminate raw waters or when aggressive waters dissolve the
cement in asbestos cement water mains.
Sometimes asbestos fibers are found in raw waters because natural
weathering action causes disintegration and erosion of rock formations that
contain asbestos. This occurs in the Cascade Mountains, for example, and is
the reason for the high concentration of asbestos fibers in the Tolt Reser-
voir.
Human activity can result in high concentrations of asbestos fibers in
water. The Reserve Mining Company's disposal of about 67,000 tons per day
of tailings into Lake Superior caused the concentration of amphibole fibers
in the lake water at Duluth to range from tens of millions per liter to
about one billion fibers per liter during severe storms. Liquid wastes dis-
charged from industries that manufacture products that contain asbestos may
be sources of contamination. An assessment of sources of environmental con-
tamination has been published by EPA (8).
Another source of asbestos fibers in drinking water is the pickup of
asbestos in water that flows through asbestos-cement pipes. In 1974 H. L.
Olson wrote, "Water flowing through asbestos-cement pipe does not increase
-------
the level of fiber content significantly" (9). Olson's statement is correct
in some cases, but unfortunately in other instances it is wrong. Buelow
et. al., in a paper presented at the 1979 Annual Conference of the American
Water Works Association state, "... asbestos-cement pipe behaves much like
other piping materials, excepting plastic, that are in common use for distri-
bution of drinking water. If aggressive conditions toward the piping material
exist, the pipe will corrode and deteriorate. If aggressive conditions do not
exist toward the piping material the pipe will not corrode and deteriorate"
(10).
Water quality conditions that cause deterioration of asbestos-cement
pipe and conditions that do not cause the deterioration are not completely
understood at the present time (1979). Further studies on this problem are
needed. Buelow et. al. gave a list of conditions indicative of situations in
which water is not attacking A/C pipe, and also a list of conditions in which
the water is probably attacking A/C pipe. Among the signs of likely attack
was inlet water screens at coin operated laundries becoming plugged with
asbestos fibers.
Water filtration often occurs before water is pumped into a water
utility's distribution system. Although filtration would not improve a water
quality problem caused by deterioration of A/C pipe, overall water treatment
practice can influence the extent to which a water attacks pipes in the dis-
tribution system. The influence can be adverse if addition of alum and
chlorine exhausts most of the natural alkalinity in a water and drastically
reduces the pH. The treatment influence can be positive, however, if the
utility recognizes the importance of corrosion control and adjusts the
quality of the treated water so that it does not attack pipes in the distri-
bution system.
This report does not deal with corrosion control, but it is mentioned
here so that filtration plant operators will be aware that they should not
treat water in a way that removes asbestos at the filtration plant but then
causes the water to deteriorate asbestos-cement pipe in the distribution
system.
ASBESTOS FIBERS IN RAW AND FILTERED WATERS
Since the early 1970's investigators have published data on the concen-
tration of asbestos in various raw and filtered waters. A review of some of
these data is presented in this document. For a more complete report on
asbestos in drinking waters, readers should obtain a report by Millette,
Clark, and Pansing (11). As the authors carefully explain, analytical methods
for asbestos fibers in water have been different from laboratory to labora-
tory. Also, techniques have improved since analysts first began searching
for asbestos in water. Therefore, caution must be applied by anyone who
wants to make comparisons of data between laboratories. A single example
should be sufficient to illustrate the point. During the pilot plant filtra-
tion research conducted at Duluth in 1974, the Ontario Research Foundation
reported amphibole counts in the range from less than 1 X 10" to 4 X 10"
fibers per liter (f/L) for raw Lake Superior water (Reference 12, Appendix
E). From 1977 through 1979 the Lake Superior Basin Studies Group at the
-------
University of Minnesota-Duluth generally found amphibole fiber concentrations
to vary from tens to hundreds of million per liter (13). Because the dis-
charge of tailings occurred in 1974 and also in recent years, the most likely
explanation for the difference in results would be different fiber recoveries
by the methods used in 1974 and in the monitoring done since 1977.
Even though comparisons of data between laboratories can cause problems,
fiber counting data for samples done within a single laboratory should be
comparable if the interval of time between counting samples is not too great.
Care should be taken when looking at 1974 vs. 1979 data, because the counting
method has been improved recently, but sample data collected in a one-year or
two-year study should be comparable if the analyst did not change or improve
methods during the study.
In their study of asbestos fibers in water, Cunningham and Pontefract
reported finding on the order of 1 to 10 X 10^ f/L in tap waters in some
Canadian towns and cities (1). They reported, "... Ottawa tap water,
drawn from the Ottawa river, had considerably fewer fibres than unfiltered
river water." Cunningham and Pontefract thus gave engineers reason to expect
that water filtration could remove the fibers.
Stimulated by the Cunningham and Pontefract paper, the Ontario Ministry
of the Environment began a program of water analysis. Samples were collected
at 22 Ontario cities in August, 1972 and analyzed by the Ontario Research
Foundation. Results of the study were published by G. H. Kay in Water and
Pollution Control in 1973 (14) and in JAWWA in 1974 (15). Of particular
interest was Kay's report that fiber count was reduced about 92 percent by
the filtration plant at Windsor, Ontario. Counts there dropped from about
20 X" 106 f/L to less than 2 X 106 f/L when the water was subjected to what
Kay called "complete treatment filtration" (14). The Windsor plant utilizes
coagulation, flocculation, sedimentation, and filtration through dual media.
In a letter to 0. J. Schmidt of Black & Veatch in 1974, Kay mentioned coagu-
lation, flocculation, and filtration as a technique to remove fibers to below
1 X 10° f/L (16).
Wigle published some data on raw and filtered municipal water supplies
in the communities of Asbestos and Drummondville in Quebec (17). At one loca-
tion the fiber count was reduced from 1,200 X 106 f/L to 200 X 106 f/L by
filtration. At another the reduction was from 680 X 106 f/L to 1.1 X 10"
f/L. Wigle was a medical doctor writing about cancer mortality in relation to
water supplies, and he included no engineering information in his paper other
than whether or not a municipal water was filtered.
In the United States, Sargent should be recognized for his early concern
about levels of asbestos in Vermont drinking waters (18). Unfortunately,
Sargent's fiber counts were made by optical microscope at a 450-power magni-
fication, and that method is no longer used for water analysis.
In an epidemiological study made to determine if a link existed between
cancer incidence and exposure to asbestos in drinking water, Kanarek reported
the results of an extensive sampling and analysis program carried out in the
San Francisco Bay area (19). Kanarek included data on filtered and unfiltered
-------
waters, as well as on water quality at many locations in the distribution
systems in the Bay area. Some of his data on filter plant performance are
presented and discussed later in .this report.
A long-term electron microscope investigation of water quality was
carried out at the Water Purification Laboratory at Chicago. McMillan,
Stout, and Willey reported on asbestos in raw and treated water at Chicago,
and presented a graph showing monthly average raw and treated water fiber
counts for the July, 1974 - December, 1975 time period (20). The Chicago
water filtration plants employ coagulation, flocculation, sedimentation and
filtration. The authors stated that the plants "... remove 70 to 90% of
the asbestos fibers present in the raw water of Lake Michigan" (20).
The monitoring program begun in 1974 has been continued. Fiber counting
data for 1976 through 1978 were provided by McMillan, and these are presented
and discussed later in this report.
TREATMENT STUDIES REPORTED BY OTHERS
Canadian Research
Much of the work on water filtration for asbestos removal has been done
with the assistance of the Environmental Protection Agency. Of the studies
done independently of EPA, work in Canada has probably been more extensive
than that done anywhere else. The Ontario Ministry of the Environment and
Canada Centre for Inland Waters sponsored research studies over a period of
years.
In the first research reported upon, Lawrence et. al. performed labora-
tory tests of flocculation and sand filtration (21). They reported 99.8
percent fiber removal when treating a lake water containing 12.3 X 10° f/L.
Later Lawrence and Zimmerman reported that with respect to fiber removal, an
optimum dose for polymer existed (22). Either an overdose or an underdose of
polymer resulted in a higher residual fiber count in water that had been
flocculated and filtered. An optimum dose for polymers has been reported by
other researchers who assessed filter behavior in terms of turbidity removal
and particle counting (23-26).
Lawrence and Zimmerman reported on use of mixed media filtration and
diatomaceous earth (DE) filtration in a study of asbestos mine effluent water
treatment (27). They reported that although mixed media filtration could
reduce the asbestos fiber concentration by a factor of 40 to 100, DE filtra-
tion could reduce the asbestos fiber concentration by a factor of 10-^.
Diatomite coated with aluminum hydroxide according to a Johns-Manville tech-
nique was even more effective, reducing the fiber count from 10^ f/L to less
than 105 f/L.
The Ontario Ministry of the Environment set up a pilot plant in its
Research Test Facility in Toronto in order to evaluate the efficacy of water
treatment processes for asbestos fiber removal. The authors reported on
problems experienced in diluting chrysotile and adding it to the pilot plant
flow stream (28) and on results of the pilot plant study (29). The pilot
-------
plant was similar to conventional plants, having rapid mix, flocculation,
sedimentation, and filtration. The plant operated at the rate of 0.4 L/s
(6.3 gpm). The plant had three 14 cm inside diameter filter columns, two of
which had dual media, while the third filter contained only sand. The authors
reported that optimized conventional treatment could significantly reduce
asbestos levels in potable water. In the most successful experiments fiber
reductions through the complete treatment train ranged from 95 to 99.3 percent.
Foley and Missingham presented data showing that in one experiment at the
test facility, 99.5 percent reduction of fibers (from 7.8 X 106 to 0.038 X
106 f/L) had taken place (30).
Everett, Washington
Water treatment pilot plant research was conducted at Everett, Washing-
ton in 1977. Filter performance evaluation was usually based upon turbidity
or particle counting data. Watkins, Ryder, and Persich did present a limited
amount of asbestos fiber data, however (31). The data are shown in Table 1.
TABLE 1. ASBESTOS COUNT CORRELATIONS - EVERETT PILOT PLANT
Asbestos Count
Run*
106 Fibers/Liter Turbidity
Sample Chrysotile Amphibole ntu l-5um 5-60um l-60jim
Particle Count
106/liter
Conventional
Treatment
#3
Direct
Filtration
#3
In-Line -
Filtration
n
Raw
Eff.
Raw
Eff.
Raw
Eff.
143
0.07
71.6
1.12
35.9
0.39
4.68
<0.01
1.7
<0.01
2.12
<0.01
1.05
0.06
0.73
0.05
0.60
0.15
12.11
0.44
7.08
0.037
5.71
0.17
1.24
0.06
0.38
0.003
0.23
0.01
13.35
0.50
7.46
0.040
5.94
0.18
*Based on a 5 GPM/Sq. Ft. Filter loading
The pilot plant, as described in detail by the authors, could operate
with conventional treatment (rapid mix, tapered flocculation, sedimentation,
rapid sand filtration), direct filtration (sedimentation omitted),and in-line
filtration (both flocculation and sedimentation omitted). The pilot plant
was rated at 1.6 gpm. Conventional treatment and direct filtration (with
flocculation) were processes capable of producing filtered water with a
turbidity of 0.10 ntu or less. Filtration preceeded by rapid mix only
10
-------
(in-line filtration) did not attain the 0.10 ntu level. The filtration
processes were effective for removing both amphibole and chrysotile fibers.
FIBER REMOVAL BY FILTRATION
The literature reviewed, when considered as a whole, suggests that water
filtration processes can be quite effective for reducing the concentrations
of asbestos fibers under certain circumstances. On the other hand, some
results show that fiber removal efficiency was not especially good. A basic
premise of this report is that reasons do exist for differences in water
filtration efficacy. Finding the reasons and explanations for differences
in filter performance is an objective that will be pursued in further sections
of this report.
Research on filtration for asbestos fiber removal has been sponsored by
EPA since 1973. The following portion of this report describes such research
on a location-by-location basis. Data are available for Duluth, Minnesota;
other communities on the north shore of Lake Superior; Seattle, Washington;
Philadelphia, Pennsylvania; and the San Francisco Bay area (through Kanarek's
study).
11
-------
SECTION 5
WATER UTILITIES ON LAKE SUPERIOR'S NORTH SHORE
DULUTH
In June, 1973 the Environmental Protection Agency announced that its
Environmental Research Laboratory in Duluth had found asbestos fibers in the
drinking water of Duluth. At the time the water was disinfected and fluori-
dated but not filtered. In the summer of 1973 EPA began a program of research
on asbestos fibers in drinking water, and activity continues to the present
time. The first research, performed in Cincinnati with Lake Superior water
that had been shipped from Duluth via air freight, indicated that dual media
filtration could reduce the concentration of larger amphibole fibers by as
much as ninety percent. However, nearly all of the asbestos analysis was done
by light microscopy at a magnification factor of 450. This technique was soon
shown to be inadequate for such work, so details of the Cincinnati research
were not published.
During the summer of 1973, the U. S. Army Corps of Engineers conducted a
brief program of filtration research at Duluth. This work was described by
Schmitt, Lindsten and Shannon (32). They reported that polymer coagulation
followed by diatomaceous earth filtration was effective for removal of
asbestos. The authors reported that all the asbestos was removed in one run.
The analytical method in 1973 probably has a practical detection limit of 0.05
to 0.1 X 10" f/L, however, so 100 percent removal of fibers was not actually
demonstrated. Nevertheless, DE filtration on a scale of 7 gpm appeared
promising.
Pilot Plant Research
In the fall and early winter of 1973 an interagency agreement for studies
of the problem was formulated and signed by the EPA and the U.S. Army Corps of
Engineers. Under this agreement, EPA funded the pilot plant research on
asbestos fiber removal and the Corps funded a study of alternative water
sources and sites for construction of a filtration plant or plants for the
Duluth-Cloquet-Superior area.
Black & Veatch conducted the pilot plant research at the Lakewood pumping
station in Duluth, Minn., with the assistance of the Department of Water and
Gas of the City of Duluth during the period from April through September,
1974. In this time, a total of 227 granular media and 228 diatomaceous earth
(DE) filter runs were performed.
12
-------
Pilot Plant Equipment
The apparatus used in the research has been described in Appendix B of
the report on the project (12). Two types of filters granular media and
DE were used. All units were situated in the Lakewood pump station. Total
water flow through individual filter systems generally ranged from 10 to 20
gpm.
Two granular filters were employed. Both units were package plants* with
4.0 sq. ft. of filter surface. Equipment variations with these units included
use of dual media; mixed media (tri-media); no settling; tube settlers;
single-stage rapid mix and two-stage rapid mix with propeller mixers; two-
stage and three-stage rapid mix with in-line mixers; alum or ferric chloride
as the primary coagulant; anionic, cationic, and non-ionic polymers; and
filtration rates of 2 - 7 gpm/sq ft.
Two kinds of DE filter systems were employed. In the pressure filtra-
tion unit' the clear Lake Superior water merely was passed through the pre-
treatment portions of the filter on its way to the pressure filter. The
filter had two pressure vessels, each containing six cylindrical septa. Total
filter-surface area for one pressure vessel was 10.0 sq. ft. After the
filter septum was precoated, body feed could be added dry or in slurry form.
The gravity or vacuum DE filter unit consisted of an open rectangular
tank with flat septa. The driving force for filtration was the difference
between atmospheric pressure and the pressure at the pump intake on the
effluent side of the filter. Filter surface was also 10.0 sq. ft. on this
unit. Body feed could be added dry or in slurry form.
Both kinds of DE filters were operated in various ways to evaluate con-
ditioning of DE with alum, cationic polymer, and anionic polymer. On some
runs, a cationic polymer was added to the raw water. Single-step vs two-
step precoat was studied. Conditioned DE was used in precoat situations as
well as for body feed. Various grades of DE, from fine to coarse, were
evaluated.
Study Results
The two principal objectives of the research were (1) to obtain infor-
mation on asbestos fiber removal and (2) to operate pilot plants in such a
way as to generate data for engineering design and cost estimates. These
objectives were accomplished. Fiber removal results are discussed in the
following sections of this report. Cost estimates based on the pilot plant
work are not presented because actual construction cost data are now avail-
able.
*Neptune - Microfloc WB-27
TErdlator, The U.S. Army water filtration system developed by the Engrg.
R&D Lab. at Ft. Belvoir, Va.
#A product of BIF Industries
13
-------
Amphibole dataAnalysis of the data showed that certain filtration
techniques were effective for amphibole removal as measured by x-ray diffrac-
tion at the National Water Quality Laboratory (NWQL, now the Environmental
Research Laboratory) and by transmission electron microscopy. Data are sum-
marized in Tables 2-4.
TABLE 2. AMPHIBOLE MASS AND FIBER REMOVAL BY DUAL MEDIA FILTRATION -
DULUTH PILOT PLANT
NWQL Ontario Research Foundation
Amphibole Mass Amphibole Fibers
Number of Samples Number of Samples
Treatment Technique
Total _<.005
Analyzed mg/L
Total At or
Analyzed Near BDL*
Filtration without sedimentation
Alum & nonionic polymer
2 gpm/sq ft
Alum & nonionic polymer'
4 gpm/sq ft
Alum & nonionic polymer'
6-8 gpm/sq ft
FeClo and cationic polymer
4 gpm/sq ft
Filtration with sedimentation
tube settlers, 4 gpm/sq ft
Alum and nonionic polymer'
o and nonionic polymer
12
2
3
4
12
1
3
10
3
2
12
2
2
9
2
2
12
2
*At or near BDL (below detectable limits) is defined as _< 0.04 X 106 f/L
'985 N, Nonionic polymer by American Cyanamid, Wayne, N.J.
#N-17, Dow Chemical Co., Midland, Mich.
§C-31, Dow Chemical Co., Midland, Mich.
From Black & Veatch (12)
14
-------
TABLE 3. AMPHIBOLE MASS AND FIBER REMOVAL BY MIXED MEDIA FILTRATION -
DULUTH PILOT PLANT
NWQL Ontario Research Foundation
Amphibole Mass Amphibole Fibers
Number of Samples Number of Samples
Total
Treatment Technique Analyzed
Chemicals added to mixing chamber:
Alum and nonionic polymer,'
4 gpm/sq ft
Chemicals added to two flash mixers:
Alum and nonionic polymer.
4 gpm/sq ft
Alum and anionic'^ and cationic^
polymers, 4 gpm/sq ft
Alum and anionic* and nonionic'
polymers, 4 gpm/sq ft
Alum and nonionic polymer.
2 gpm/sq ft
In-line mixers:
Alum and nonionic polymer.
4 gpm/sq ft
Alum and nonionic' and anionic*
polymers, 4 gpm/sq ft
Alum and anionic'^ and cationic^
polymers, 4 gpm/sq ft
Alum and nonionic polymer'
6 gpm/sq ft
Alum and cationic polymer, s
4 gpm/sq ft
1
5
2
2
1
1
2
1
2
1
<.005
mg/L
1
5
2
2
1
1
2
1
2
1
Total At or
Analyzed Near BDL*
1
5
2
2
1
1
2
1
2
1
1
5
2
2
1
1
2
1
2
1
*At or near BDL (below detectable limits) is defined as jC0.04 X 106 f/L.
^985 N, Nonionic polymer by American Cyanamid, Wayne, N.J.
*A-23, Dow Chemical Co., Midland, Mich.
&C-31, Dow Chemical Co., Midland, Mich.
From Black & Veatch (12)
15
-------
TABLE 4. AMPHIBOLE MASS AND FIBER REMOVAL BY DIATOMITE FILTRATION -
DULUTH PILOT PLANT
NWQL Ontario Research Foundation
Amphibole Mass Amphibole Fibers
Number of Samples Number of Samples
Treatment Technique
Total
Analyzed
<.005
mg/L
Total
Analyzed
At or
Near BDL*
Pressure filtration;two-step precoat,
1 gpm/sq ft
Anionic polymer' to second step of 1
precoat; alum & soda ash to body
feed
Alum & soda ash to second step of 6
precoat
Alum & soda ash to second step of 3
precoat
and to body feed
Cationic polymer' to raw water 5
Alum & soda ash to second step of 3
JLr
precoat; cationic polymer" to
raw water
Vacuum filtration;one-step precoat,
1 gpm/sq ft:
Anionic polymer' to precoat 2
Vacuum filtration;two-step precoat,
1 gpm/sq ft:
Anionic polymer' to total precoat 0
Anionic polymer^ to second step of 1
precoat
Anionic polymer' to second step of 5
precoat; alum & soda ash to
body feed
Alum & soda ash to second step of 2
precoat
Alum & soda ash to second step of 2
precoat
and to body feed
Cationic polymer^ to raw water 1
Alum & soda ash to second step of 4
ir
precoat; cationic polymer'' to
raw water
6
3
5
3
0
0
2
2
1
4
6
3
5
3
1
1
3
4
3
3
5
3
1
1
0
2
3
4
*At or near BDL (below detectable limits) is defined as < 0.04 X 106 f/L.
tA-
#
A-23, Dow Chemical Co., Midland, Mich.
Catfloc-B, Calgon Corp., Pittsburgh, Pa.
From Logsdon and Symons (33)
16
-------
Data in Tables 2-4 indicate that amphibole asbestos fibers can be removed
readily by filtration. Additional evidence to confirm the efficacy of filtra-
tion was found in the amphibole mass data. Many of the filter runs that were
sampled contained 0.005 mg/L or less in the filter effluent. Based on the
amphibole mass concentration in the raw water, this represented amphibole mass
reductions of tenfold to fortyfold. Treatment of the raw water with alum and
a nonionic polymer was considered to be the most effective for amphibole fiber
removal by granular media filtration. Techniques most effective for removing
amphiboles by diatomaceous earth filtration involved conditioning the DE
filter aid with alum or a polymer, or conditioning the raw water with cationic
polymer. Pressure DE filtration was more effective than vacuum DE filtration,
probably because air bubbles would form and collapse within the diatomite
filter cake when the cold, highly oxygenated Lake Superior water experienced a
pressure drop to less than one atmosphere in the pilot plant vacuum DE filter.
In late 1974 personnel from the Johns-Manville Research and Engineering
Center performed DE filtration tests at Duluth. The results are described in
a Johns-Manville report (34) that was prepared in response to the Interim
Report on Water Supply for the Duluth-Cloquet-Superior Area (35). The Johns-
Manville report indicates that diatomaceous earth filtration is very effective
for asbestos fiber removal. Analysis by optical phase contrast microscopy at
1000 X showed that 99.80 and 99.93 percent removal of amphibole fibers with a
length of 5 urn or greater. These results, although limited in scope, agree
with the other results from Duluth and elsewhere.
Reevaluation of chrysotile dataBoth the project report prepared by
Black & Veatch (12) and the analysis of fiber data by Logsdon and Symons (33)
included data on chrysotile fibers, as well as on amphibole fibers. The data
were reported in good faith by the electron microscope laboratory, and
although the chrysotile counts were questioned at the time by EPA personnel
working in Duluth, filtration researchers did not have enough fiber data from
other sources to decide to discard or disregard the chrysotile data.
A more recent analysis of chrysotile data suggests they should be dis-
regarded. Originally data were evaluated in groups according to the type of
treatment, including treatment chemicals used. The basis for evaluation was
to consider the number of filtered water samples with chrysotile equal to or
less than 0.04 X 10^ f/L. Raw water fiber counts were not considered. A
new analysis of the chrysotile data takes into account recent Duluth and
Seattle experience. The data in Table 5 are grouped according to whether
or not filtered fiber counts are less than raw water fiber counts, with only
those samples having filtered water turbidity of 0.10 ntu or less included.
Seattle results showed consistent removal of chrysotile at this turbidity
level but these data did not.
The hypothesis for analysis is that all data are random and filtered
water fiber counts would be expected to be less than raw water counts for 50
percent of the observations. At the 0.05 level the hypothesis can not be
rejected for any of pilot filter units employed at Duluth.
17
-------
TABLE 5. COMPARISON OF RAW AND FILTERED WATER CHRYSOTILE DATA FOR DATA
PAIRS WITH FILTERED WATER TURBIDITY EQUAL TO OR LESS THAN 0.10 ntu -
DULUTH PILOT PLANT
Treatment Data Comparison of Chrysotile
Unit Pairs Fiber Count Data
raw exceeds raw less raw equals
filtered than filtered filtered
MM1 15 11 40
(granular media)
MM2 25 13 84
(granular media)
vacuum DE 3 0 30
pressure DE 17 8 7 2
In addition to the statistical analysis, other EM data support disre-
garding the earlier Lake Superior chrysotile data. Chrysotile fibers are
seldom found by the University of Minnesota-Duluth laboratory (13) and by the
EPA laboratory in Duluth. Therefore, chrysotile data from the Duluth pilot
plant study are not presented in this report.
A number of publications resulted from the Duluth pilot plant study.
Conclusions related to design and cost factors have been given by Robinson
et. al. (36). A paper on DE filtration optimization has been presented by
Baumann (37). EPA published the results of the pilot plant study (12).
Logsdon and Symons published an analysis of the pilot plant performance
data (33).
Filtration Plant Design and Construction
After the pilot plant research was completed, the City of Duluth hired
Black & Veatch to prepare an engineering report on the design of a filtration ,
plant. The consultant recommended building a gravity, granular media filtra-
tion plant with mixed media (tri-media) filters. In February, 1975 the City
Council approved the report and authorized Black & Veatch to prepare plans and
specifications of a 30 mgd plant (38). This was done during the spring and
summer of 1975.
The City of Duluth worked through the Minnesota congressional delegation
to obtain a demonstration grant for water filtration. The efforts were suc-
cessful, and EPA awarded a $4,000,000 demonstration grant on October 6, 1975
(retroactive to June 27 of that year). This has been described by Peterson
(39).
Construction progress was unusually rapid. Ground was broken in August,
18
-------
1975. The plant dedication ceremony was held in 1976 at the end of November.
Construction management techniques have been described by both Patton (38) and
Peterson (39). The filtration plant, as dedicated in 1976, was a direct fil-
tration plant providing rapid mixing and filtration. A basin used as a
chlorine contact basin until November, 1976 was converted for use as a floccu-
lation and sedimentation basin. The rebuilt basin was completed and put into
use in the process flow stream in June, 1977. Patton described the plant in
detail (38). A flow diagram for the plant, shown in Figure 1, reveals that an
unusual amount of process flexibility exists at Duluth. Rapid mixing can be
accomplished in three stages, chemical addition can be done at numerous loca-
tions, and any or all of the unit processes prior to filtration - rapid mixing,
flocculation, and sedimentation - can be bypassed. Hydraulic information on
the plant, as given by Patton, can be found in Table 6.
TABLE 6. PLANT HYDRAULIC INFORMATION - DULUTH
Plant Design Plant Overload
Rate (30 mgd) Rate (36 mgd)
Rapid mix chambers
Detention each, seconds 30 25
G value sec"1 500
Flocculation Facilities
Detention, min 40 33
Sedimentation Facilities
Detention, min 140 117
Surface loading rate, gpm/sf 1.1 1.3
Weir overflow rate, gal/lin ft/day 20,000 24,000
Filters
Filtration rate, gpm/sf 4.9 5.9
Wash Water
Design rise rate, in/min 30
Duration, min 10 -
Recovery storage, gal. 240,000
Wash water recovery facilities
Detention, min 285 240
Surface loading rate, gpm/sf 0.5 0.6
Flocculation zone, min 30 25
From Patton (38)
The plant has four filters, each with 1067 square feet of surface area.
Filtration rate is 4.9 gpm/sf at 30 mgd. Three filters have mixed media. For
purposes of comparison, one filter has dual media. Surface wash was provided
for all filters. Media specifications are given in Table 7.
19
-------
Rapid mix units
Lake Superior
Chemicals
Existing Lakewood
pumping station
Floccufation Sedimentation
KT~
Wash water Chemicals
settling basin f Transfer
pumps
Wash water
recovery
pumps
Wash water
storage basin
Siudge pumps
Sludge storage
lagoons
Chemicals
To finished
water
reservoir
Figure 1. Flow diagram for filtration plant at Duluth, showing treatment
train options.
From Logsdon (40)
-------
TABLE 7. DULUTH FILTER MEDIA
Depth
Size
Uniformity coefficient
Mixed Media
Coal
Sand
Garnet
Dual Media
16.5"
9"
4.5"
1.0-1.1 mm
0.42-0.45 mm
0.21-0.32 mm
1.7 max
1.35 to 1.70
1.8 max
Coal
Sand
21"
9"
1.0-1.1 mm
0.42-0.45 mm
1 . 7 max
1.35 to 1.70
Operating Results
Primary Coagulant, pH Control, and Polymers
As a result of the pilot plant research, the plant was designed to use
liquid aluminum sulfate as a primary coagulant with liquid sodium hydroxide
for subsequent pH control, nonionic polymer for amphibole fiber removal and
anionic polymer for chrysotile fiber removal.
Because EM analysis of water samples indicated no chrysotile fibers in
raw or finished water, Duluth Filtration Plant personnel decided to discon-
tinue the use of anionic polymers. After successfully eliminating the anionic
polymer from pilot plant runs, it was eliminated on a trial basis in the main
filter plant.
As a result of the pilot plant and full scale studies, anionic polymer is
no longer used at the Filtration Plant. The water is conditioned with alum,
and the nonionic polymer strengthens the floe so that turbidity breakthrough
is less likely to occur.
Turbidity is monitored continuously at each filter. The continuous flow
turbidimeters are used to detect changes in filter effluent quality. A labor-
atory turbidimeter is used to calibrate the flow-through turbidimeters and to
obtain the daily filter plant effluent turbidity value that must be recorded
to satisfy the Interim Primary Drinking Water Regulations.
Water Quality
MonitoringThe Duluth filtration plant was built to remove asbestos
fibers from Duluth drinking water. The water quality data collected since
January, 1977 show that the plant performance has exceeded expectations. The
operating goal at Duluth has been to produce the best possible filtered water
quality. The plant consistently produces filtered water turbidities in the
0.04 to 0.06 ntu range, as shown in Figure 2. Turbidity of filtered water is
almost always below 0.10 ntu except when changes in chemical feed or filtra-
tion rate cause higher turbidity. Raw water turbidity in Lake Superior at the
Duluth intake was 1 ntu or less about ninety percent of the time from 1952 to
1972 (37).
21
-------
100 -
10
«-»
C
fc 1.0
Q
£
<- ?
10 O.li
0.01
J
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 \ 1 1 1 1 1 1 1 1 1 1 ' 1
8° ° o
RAW WATER
0 0 Q-PCPV, ^tP^O>P 0°0 Ql 0 o
#"P VP000^0000 ^ ^o oo o ° o o
' ] °*ap I
* 1 1 *-. 1 -^-
*^\/^^w ^»v»«~«»*»~X~»«^*^«A *?*£* J\^ +*r**~
1 1 P I1. Iz
!i!i II il
1 2 1 L_ 1 -> ^ 1 1 .ft-
i°ii sg i s
ia K lo "" z| IE
!s!= K ill is
RAPID MIX, FLOCCULATION |>< |Q<| S
SEDIMENTATION, |S 1 a fE 19
FILTRATION g |<^ |^
11 i i r
, i i i i > I I 1 1 1 1 i 1 1 1 1 1 1 J 1 ' 1
AN MAR MAY JUL SEP MOV JAN MAR MAY JUL SEP NOV JAN MAR MAY JUL
1977 1978 1979
Figure 2. Lakewood filtration plant turbidity data
-------
Raw and filtered amphibole fiber counts are shown in Figure 3. Filtered
water amphibole fiber counts have consistently been below 0.1 X 10^ f/L since
the plant operation started. Monitoring data are tabulated and presented in
Appendix A-l.
Amphibole fiber concentrations in filtered water were higher in January,
1977 than in later months. In fact, fiber counts through September, 1977
seem to be higher than subsequent counts. This trend may have occurred be-
cause as local personnel gained experience in their jobs they could improve
plant operating techniques as well as electron microscope (EM) analysis tech-
niques. Training operators in the finer points of plant operation at a
specific plant might take a number of months. Also, eliminating little prob-
lems that usually occur in a totally new plant takes time. The lowest reported
concentration for amphibole fibers decreased from 0.19 X 106 in January, 1977,
to 0.067 in February, to 0.030 in May, and to 0.010 X 106 f/L in September.
The lower concentrations were made possible by improved analytical procedures
(analyzing a greater surface area of the sample grid) and by production of
water with a fewer particulates (resulting in filtration of larger sample
volumes for EM analysis).
Because progessive improvement in operating and analytical capabilities
took place during the first year of operation, a comparison of direct filtra-
tion data and conventional treatment data should not be made with 1977 results.
Determining how much improvement in water quality was caused by flocculation
and sedimentation and how much was caused by the improvement of skills that
occurred in the first nine months is not possible.
Results during stormsDuring the five months of pilot plant studies in
1974, the raw water turbidity did not exceed 5 ntu. A key question related to
the efficacy of the filtration process when Lake Superior was turbid and
amphibole fiber counts were on the order of one billion fibers per liter. In
the fall of 1977, storms on Lake Superior caused raw water turbidity and
asbestos fiber count to increase sharply. The efficacy of the filtration plant
was tested twice when amphibole fiber counts were very high, and on both
occasions plant performance exceeded expectations and confirmed the soundness
of the plant design and the capabilities of plant operating personnel. Data
obtained during the storms are shown in Table 8.
The treatment plant effectively removed as much as 99.996 percent of the
asbestos fibers during storm conditions. Fiber removal was enhanced by
adjusting the chemical feed after raw water quality had changed. The data
show that maintaining adequate chemical dose is important to attain maximum
fiber removal and to provide assurance that the plant is very effective for
fiber removal during storms.
Distribution systemsUnfiltered water containing amphibole asbestos
fibers was pumped into the Duluth distribution system for nearly twenty years.
While the filtration plant was under construction, a program of water main
flushing was carried out. Duluth has numerous water storage reservoirs in the
distribution system. About three days' supply can be stored. This type of
distribution system was developed in past years so that drinking water could
be provided without pumping lake water when storms caused high turbidity.
23
-------
1000
100
. 10
cc
LU
CO
a 1.0
O
m
i
a.
< 0.1
0.01
RAW WATER
D
"stf3
4HfTT
n %
'^ i
i
^BL i
BBP
₯PW * X""
FILTERED WATER
RAPID MIX
FILTRATION
' 1 1 1 i
i i i i r
D°
J7p~| i j
LJLfM r~5J
r j Q_^
&
L
X
5
. 9
Q.
^
<-...,
»
RAPID MIX, FLO(
i ' i T r i 1 1 1 1 1 1 1 1 1 1 1 1 1 1
a n
^fanrp D D
PI n ^^ n n ^ n
D 4^h D n n n ^
r-rf-l ^^ d L3 n
nfiPR DCi a.
]On n
n
2
PI
<
oc |
^ 1
z
o
^
^
o
u -Z
o 2
^. **
f
5 _j
^
LL 1 Q
L -.-*.*
! _ "f
!""f" "f
:CULATION
SEDIMENTATION, FILTRATION
1 1 1 1
Q.
cc
r
"=1 D n
2
ip
Q<
2 £
4 '
CC ul
" B
T ^^^^^7^ ^r^
LEGEND: Samples below
detection limit(BDL)
are shown at the detect-
z
o
\-
cc
\-
~1
ul
X
Q
Q. _
<
CC
g
,1,1.1 ,ionl!mit-^
MAR MAY JUL
1977
SEP
IMOV JAN
MAR MAY JUL
1978
SEP
NOV JAN
MAR MAY JUL
1979
Figure 3. Lakewood filtration plant amphibole fiber data.
-------
TABLE 8. DULUTH FILTRATION PLANT PERFORMANCE DURING STORMS
Sample Raw Water Filtered Water
Date Description Turbidity Amphibole Turbidity Amphibole
ntu Fibers ntu Fibers
106 f/L 106 f/L
9/19/77 Raw water 14.0 1200
9/19/77 Filtered water 0.045 0.048
10/8/77 Raw water 8:30 pm 11.0 830
Filtered water 8:10 pm 0.05 0.12
just before chemical
feed adjustments
10/8/77 Filtered water midnight 0.045 0.029
3 1/2 hours after
chemical feed adjustment
Flushing the reservoirs has been done as a part of normal distribution
system operation. When the filter plant operation started, a substantial
reduction of fibers took place, in the reservoirs, from hundreds or tens of
millions, to less than ten million. In March, 1977 amphibole fibers were
generally less than 1 X 10^ in reservoirs. In April, 1979 amphibole fiber
counts in the distribution system were similar to amphibole fiber counts of
filtered water at the Lakewood plant. These results are summarized in Table
9.
Effects of rate changes on filtered water qualityWhen a change in fil-
tration rate occurs, the quality of filtered water can also change. At Duluth,
a number of circumstances bring about rate changes of different types. Some
are related to filter backwashing, and others are caused by plant production
changes.
Raw water pumps at Duluth have capacities of 20 mgd and 30 mgd, so the
plant operates at one or the other of those flow rates. A 50 percent rate
increase would occur if production was boosted from 20 mgd to 30 mgd (3.2 to
4.9 gpm/sf). Since the plant was built, Duluth water consumption has typi-
cally been less than 20 mgd. Thus the plant is operated on an intermittent
basis, being shut down when additional water production is not needed. Start-
ing the plant results in a rate increase of 0 to 3.2 gpm/sf. With floccula-
tion and sedimentation basins in use, filter rate changes caused by production
changes are gradual, because of the design of the sedimentation basin effluent
collectors. These changes are considerably more abrupt when the treatment
process consists of rapid mixing followed by filtration.
When one filter is removed from service for backwash, the other three
filters experience a 33 percent rate increase (3.2 to 4.3 gpm/sf) in a 30 to
60 second time period. After backwashing is completed, the clean filter is
returned to service. This causes a very rapid rate increase in the clean
filter.
25
-------
TABLE 9. AMPHIBOLE FIBER CONCENTRATIONS IN DULUTH DISTRIBUTION SYSTEM RESERVOIRS
1977
Reservoir Distance
from Plant
'A" Nearest
Endion
Woodland
Highland
Middle
W. Duluth Farthest
2/11
10 6
f/L
0.61
1.8
3.4
N.T.
9.7
4.4
3/10
106
f/L
N.T.
0.21
0.21
1.2
N.T.
0.42
3/17
10 6
f/L
0.36
0.11
0.16
1.3
0.50
N.T.
3/22 10/25
106 106
f/L
0.32
3.0
BDL(0.48)
1.3
0.50
0.48
f/L
0.15
0.27
BDL(0
BDL(0
1.2
1.4
1978
6/08
106
f/L
BDL(0.032)
N.C.
.87) 0.024
.12) N.C.
0.048
N.C.
9/19
10 6
f/L
0.22
0.11
0.17
0.13
0.095
0.21
1979
4/30
f/L
N.T.
0.13
0.09
0.04
0.02
BDL
(0.019)
N.T. - no sample taken
BDL - below detectable limits (detection limit in parentheses)
N.C. - sample could not be counted because of excessive debris
-------
Data have been collected for various types of filter rate changes. Work
on this is not yet completed. Data obtained and reported by the Duluth
Department of Water and Gas are presented in the Appendix A-2. Data used to
evaluate water quality changes were filtered water turbidities and amphibole
fiber counts. Quality of the water after filter rate changes was compared to
quality before the rate change occurred. Results in Appendix A-2 suggested
that fiber count and turbidity generally increase only slightly when a clean
filter is started after it is backwashed.
At Duluth, because all backwash water is discharged to a washwater treat-
ment plant, operating practice is to shut down the plant without washing all
filters immediately before treatment stops. On a few occasions, filters were
restarted at head losses of greater than 6 feet. This resulted in much higher
fiber counts in the filtered water for a brief period (less than on hour) on
two occasions.
Additional work
Studies of plant performance and monitoring of water quality for asbestos
fiber concentration will continue at Duluth into mid-1980, at which time the
research requirements for the demonstration grant should be met.
OTHER NORTH SHORE PLANTS
In addition to Duluth, a number of other communities along the Lake
Superior shore were affected by the taconite tailings discharge at Silver Bay.
Of these, Silver Bay and Two Harbors have built water filtration plants. Con-
struction was being completed on the filtration plant at Beaver Bay during
the summer of 1979, so no data are included for Beaver Bay.
Before comparisons of water quality at filtration plants in Duluth, Two
Harbors and Silver Bay are made, the nature of Lake Superior currents and
their effect on water quality needs to be explained. Along the north shore
of Lake Superior, in the western portion of the lake, the current in the lake
flows from the northeast, past Silver Bay, then Two Harbors, and then past
the Duluth intake. The current turns at the head of the lake at Duluth,
Minnesota and Superior, Wisconsin, and flows along the shore past northern
Wisconsin and the Upper Peninsula of Michigan. The current flows past the
Silver Bay intake before tailings from the Reserve Mining Company are dumped
into Lake Superior. The- Silver Bay intake is close enough to Reserve's flume
to be affected, but amphibole fiber counts at Silver Bay tend to be lower
than counts at Two Harbors or Duluth. Because Two Harbors is closer to
Reserve's flume than Duluth, amphibole fiber counts are higher at Two Harbors.
Raw water quality at all three plants is similar, but fiber counts differ for
the above reason.
Two Harbors
The two Harbors plant is a 2.6 mgd direct filtration plant designed by
RREM, Inc. It employs two-stage rapid mix, two-stage flocculation, and fil-
tration through mixed media. A flow diagram is shown in Figure 4. Plant
hydraulic information is given in Table 10. The plant has three filters,
27
-------
LAKE
SUPERIOR
K3
00
Alum
Chlorine Polymer
Flocculation
Filters
Fluoride
Caustic Clearwell
To
Sewer
Backwash
Detention Tank
To Distribution
System
High Service Ground Storage
Pumps and Chlorine Detention
Waste Pumping
Station
Figure 4. Flow diagram for filtration plant at Two Harbors.
-------
each with 150 square feet of surface area. Media design is given below:
Material Specification Depth
Coal Neptune - Microfloc MS - 4 16.5"
Sand Neptune - Microfloc MS - 6 9.0"
lllmenlte Neptune - Microfloc MS - 21 4.5"
The Two Harbors filtration plant design was based upon results from the
Duluth pilot plant study. The plant uses alum and a nonionic polymer to con-
dition the water for filtration. Both of these chemicals are added at the
rapid mixers. Caustic soda is added in the clearwell to adjust pH before the
water is pumped to the distribution system. Turbidity is monitored continu-
ously at each filter, and grab samples are taken periodically so that turbidity
can also be measured with a laboratory turbidimeter.
TABLE 10. PLANT HYDRAULIC INFORMATION - TWO HARBORS
Rapid mix chambers (4)
(2 in series each side)
Total detention time, minutes
Flocculation basins (4)
(2 in series each side)
Total detention time, minutes
Filters (3)
Filtration rate, gpm/sf
Wash water
Design rate, gpm/sf
Duration, minutes
Plant Design Rate, 2.6 mgd
4.2
20
15
10
The plant at Two Harbors was started early in 1978. Two Harbors obtained
some money for construction from the State of Minnesota, but an extensive
water quality monitoring program of the type undertaken at Duluth was not
possible. Therefore, only a limited amount of data on fiber removal at Two
Harbors is available. The results obtained to date are given in Table 11.
Filtered turbidity at Two Harbors generally ranges from 0.10 to 0.20 ntu, and
fiber counts typically have been from 0.4 to 2 X 10" f/L. These values are
somewhat higher than the values for Duluth and Silver Bay, although they
were taken with a different instrument and may not be comparable. Substantial
reductions in fiber count are consistently attained.
29
-------
TABLE 11. TWO HARBORS AMPHIBOLE DATA
Date
4/28/78
5/15/78
6/22/78
6/27/78
7/3/78
7/18/78
7/27/78
8/1/78
9/1/78
9/7/78
10/6/78
1/9/79
6/26/79
Lab
EPD
MDH
MDH
EPD
EPD
EPD
MDH
EPD
EPD
MDH
EPD
MDH
MDH
Raw
Turbidity
ntu
0.6
1.2
1.0
0.78
1.3
1.5
0.8
1.0
0.66
1.0
1.2
0.96
Water
Fibers
106 f/L
141
44
111
200
177
84
19
53
356
22
61
65
Filtered
Turbidity
ntu*
«
0.14
0.16
-
-
0.10
0.15
0.20
0.10
0.15
0.14
0.13
0.14
Water
Fibers
106 f/L
m_
0.43
0.80
-
-
1.6
0.56
2.3
0.45
< 0.07
1.9
0.37
0.28
Percent
Turbidity
«
88
84
-
-
93
81
80
85
85
88
85
Reduced
Fibers
_
99.0
99.3
-
-
98.1
97.1
95.7
99.9
> 99.7
96.9
99.5
EPD = EPA-Duluth
Silver^Bay
MDH = Minn. State Health Dept.
*May not compare with
Duluth and Silver Bay
The Silver Bay treatment plant was modified and put into service in its
present configuration in May, 1978. The filters, clearwell, and both low lift
and high left pumping facilities existed before the modification. New facili-
ties added in 1978 were rapid mixing and flocculating basins as well as
chemical feed and storage facilities. A schematic diagram of the modified
plant is shown in Figure 5. Design information for the plant from RREM,
Inc., .is given in Table 12. The plant has dual media filters, with 15
inches of 1.16 mm effective size coal (U.C. = 1.43) and 15 inches of 0.45 mm
effective size sand (U.C. = 1.41).
Amphibole fiber and turbidity data for the Silver Bay plant are given
in Table 13. Because of the circulation pattern in Lake Superior, raw water
fiber counts at Silver Bay are lower than counts at Two Harbors and Duluth.
The limited amount of data available indicate that the Silver Bay plant
is effective in removing amphibole fibers from water.
Future North Shore Activity
Additional monitoring for amphibole fibers is planned for Two Harbors and
Silver Bay. Monitoring will also be done at the Beaver Bay plant to determine
amphibole fiber concentrations in filtered water at the plant most influenced
by the tailings discharge from the Reserve Mining Company.
30
-------
Alum
Chlorine Polymer
LAKE
SUPERIOR
Low Service
Pumps
To Tank Truck
Waste Holding
Tank
Backwash
Clarifier
Flocculation
Backwash
Detention Tank
Fluoride
Caustic Clearwell
Filters
To Distribution
System
High Service
Pumps
Figure 5. Flow diagram for filtration plant at Silver Bay.
-------
TABLE 12. PLANT HYDRAULIC INFORMATION - SILVER BAY
Rapid mix chambers (4)
(2 in series each side)
Total detention time, minutes
Flocculation basins (4)
(2 in series each side)
Total detention time, minutes
Filters (2)
Filtration rate, gpm/sf
Wash water
Design rate, gpm/sf
Duration, minutes
Plant Design Rate, 2.3 mgd
8.1
40.5
2.6
7.4
7
TABLE 13. SILVER BAY AMPHIBOLE DATA
Date
6/22/78
7/27/78
9/5/78
1/9/79
6/26/79
Lab
MDH
MDH
MDH
MDH
MDH
Raw
Turbidity
ntu
0.41
0.41
0.32
-
0.28
Water
Fibers
106 f/L
25.3
8.4
1.8
-
14.3
Filtered
Turbidity
ntu
0.05
0.05
0.07
0.05
0.06
Water
Fibers
106 f/L
0.18
< 0.13
< 0.13
0.18
0.33
Percent Reduced
Turbidity Fibers
88 99.29
88 > 98.4
78 > 92.8
_ _
79 97.7
32
-------
SECTION 6
SEATTLE PILOT PLANT STUDY
In 1975 amphibole and chrysotile fibers were found in the South Fork Tolt
Reservoir, which supplies up to 100 mgd to a portion of the Seattle metropoli-
tan area. The clear, clean mountain water is distributed to consumers after
screening, disinfection and fluoridation. After the fibers were discovered,
the Seattle Water Department applied for and received a research grant to
study filtration of the Tolt water. Results of this work were summarized by
Kirmeyer et. al. (41). The following section is adapted from Kirmeyer1s
report (42).
DESCRIPTION OF STUDY
During the period from January 1977 through September 1978, a small
scale water filtration plant was operated at the Regulating Building on the
Tolt supply. The primary goal of the pilot filtration study was to investi-
gate various filtration systems to remove amphibole and chrysotile asbestos
from the raw Tolt water. In-line, direct and conventional filtration pro-
cesses were investigated. Various mixing intensities were tested with dif-
ferent chemical coagulants and dosages. Several granular medias were com-
pared at filter loading rates up to 10 gpm/ft .
Study Objectives
The formal objective of the Seattle research effort was to determine the
most feasible method for reducing asbestos count in the Tolt water. The scope
of work included:
1. Developing methods to improve chrysotile removal by the use of
three-stage mixing in combination with three conditioning chemicals
in granular media filtration;
2. Optimizing asbestos removal using anionic, cationic and nonionic
polymers;
3. Conducting filtration experiments at 5-6 gpm/ft^ with granular
media filters;
4. Determining the effect of mixing intensity on filtration and
comparing back mixing with in-line mixing techniques;
5. Developing an operating tool which indicates quickly and econom-
ically if asbestos removal is occurring;
6. Collecting design data and developing cost information; and
7. Confirming amphibole results gathered during the Duluth asbestos
removal study (12).
33
-------
Once the grant was approved, a literature review of the properties of
asbestos fibers and methods for their removal was conducted. CH2M/Hill Con-
sulting Engineers was awarded a contract to help with the collection of design
data and cost information and the University of Washington agreed to perform
the asbestos analysis. A study plan was developed, bench scale filtration
studies were conducted and pilot testing was then undertaken.
Elements of the Study
The study consisted of the following six primary elements that included
various secondary topics.
Water Quality
1. Seasonal Variation of Water Quality Parameters.
2. Relationship of Raw Water Turbidity with Asbestos Counts.
3. Asbestos Size Distribution.
Methods of Analysis
1. Comparison of Asbestos Analysis Methods.
2. Particle Counting Techniques.
3. Development of Turbidimeter as Operational Tool.
Pretreatment for Filtration
1. Effectiveness of Alum, Ferric Salts and Polymers.
2. Mixing Requirements Including Flocculation.
Filtration Process
1. Comparison of Granular Media Filters.
2.' Direct Filtration Techniques.
3. Filter Loading Rates Up to 10 gpm/ft2.
4. Rate of Headloss Build-Up.
5'. Air Binding Problems.
Backwash Water
1. Sludge Production.
2. Sludge Settling Characteristics.
3. Sludge Treatment Train.
Engineering Analysis
1. Preferred Techniques for Asbestos Removal.
2. Development of Design and Operating Criteria
3. Cost Estimates.
MATERIALS AND METHODS
Pilot Plant Equipment
The pilot study was conducted On a 20 gallon per minute (gpm) (WB-27)
package plant, manufactured by Neptune Microfloc, Inc. The unit was
installed in the chlorine room at the Tolt Regulating Basin. Because the
unit was designed to operate as a conventional filtration plant, modifications
34
-------
were made to the basic unit to enable several different treatment trains to be
tested. Because of a serious air binding problem which developed in the
filter media due to insufficient water over the top of the filter, three addi-
tional filter columns were constructed and operated in parallel with the
package plant filter.
The unit processes on the package plant included a hydraulic rapid mix
chamber, a flocculation basin with picket flocculators, tube settlers and a
granular media filter.
Modifications of Equipment
Because the unit was to be used as a pilot plant, the manufacturer pro-
vided a variable speed flocculator so that various mixing intensities could be
investigated. Several other modifications were made on the unit by the
Seattle Water Department. Additional piping was installed to enable any unit
process to be by-passed or eliminated simply by operating a valve or removing
a pipe or fire hose. Figure 6 illustrates the possible flow schematics and
Table 14 lists pertinent dimensions.
Three 2-inch static mixers were installed with the unit. Mechanical
mixers and 55-gallon drums were also part of the flexible treatment train.
Water would flow under pressure into the first drum and then could discharge
by gravity into a second drum or into any other unit process in the treat-
ment train.
A flow controller, which would maintain a preset flow rate regardless of
incoming line pressure, was installed on the raw water pipe and an effluent
meter was placed on the finished water line. These devices provided assur-
ance that a constant flow rate entered the plant and was discharged from the
filter.
Two turbidimeters were installed, one on the raw water and one on the
finished water. These meters fed information to a strip chart recorder to
insure that a permanent record of turbidity was kept. Bubble traps were
later installed on each turbidimeter to prevent air bubbles from giving false
turbidity readings. Filtered water was discharged to an 8,000-gallon storage
tank with an overflow near the top of the tank.
Early in the pilot studies, air binding in the package plant filter
became a serious problem. Apparently, the pressure in the filter would drop
below one atmosphere and dissolved gases, which were present in the raw
water, would come out of solution and plug the voids in the filter media.
This caused two problems: (1) rapid build-up of headloss as the filter run
progressed, and (2) loss of anthracite media during filter backwash because
of air bubble release at the start of backwash.
To remedy this situation, three 4 1/4-inch diameter filter columns were
installed in the basement of the Regulating Building (see Figure 7). A small
stream of water was diverted from the plant, above the filter media, to the
filter columns to maintain a column of water about 10 ft high over the top of
the filter media. Maintaining a positive head throughout the column filter
media held the gases in solution and prevented the air binding problem.
35
-------
OJ
RECORDING
TURBIDIMETER
TOLT
RAW
WATER
FLOW
CONTROLLER
A 8 C D CHEMICAL
E'F'G'H' = ADDITION
' ' ' POINTS
HYDRAULIC
RAPID MIX
MECHANICAL BACK MIXERS
M
^
STATIC MIXERS
1X
n
TUBE
SETTLERS
G
FLOG
CHAMBER
H
GRANULAR
MEDIA
FILTER
-^ SEE FIGURE 7
RECORDING
TURBIDIMETER
FINISHED
VATER
TO STORAGE
TANK
From Kirmeyer (42)
Figure 6. Pilot plant flow schematic Seattle.
-------
TABLE 14. DIMENSIONS OF UNIT PROCESSES - SEATTLE PILOT PLANT
Unit Process
Hydraulic
Rapid Mixing
Chamber
Mechanical
Back Mixing
Chamber
Static Mixing
u> Pipe
Flocculation
Chamber
Backwash
Settling
Chamber
Finished
Water Storage
Tank
Length Width or Diameter Depth
62.9 cm 27.3 cm 165.1 cm
24-3/4 in 10-3/4 in 65 in
58.4 cm 71 cm
N/A
23 in 28 in
4 5 . 7 cm 5 . 1 cm
N/A
18 in 2 in
62.9 cm 57.1 cm 165.1 cm
24-3/4 in 22-1/2 in 65 in
6.1m 1.2m 0.53m
20 ft 4 ft 1.75 ft
3.1m 2 . 13 m
N/A
10 ft 7 ft
Volume
0.28 m3
10 ft3
0.19 m3
6.7 ft3
933 cm3
0.033 ft3
0.59 m3
20.9 ft3
3.88 m3
140 ft3
16.1 m3
549.5 ft3
Detention Time*
4.7 min
3.1 min
0.92 sec
9.8 min
N/A
N/A
*Theoretical detention time at 60.5 1/min (16 gpm).
cm = centimeter N/A = not applicable
cm3 = cubic centimeter gal = gallon
m = meter ft3 = cubic foot
m-
cubic meter
in = inch
min = minute
sec = second
From Kirmeyer (42)
-------
FROM FIGURE
BV
FLOG
CHAMBER
'PASS
i
1
WATERBOY
FILTER
CHAMBER
/FLOW
I METER
V '
GROUND FLOOR
BASEMENT" ~
T
DIVERSION
CHEMICALLY TREATED
UNFILTERED WATER
GRANULAR MEDIA FILTER COLUMNS
FINISHED WATER
T
FSNISHED WATER
Figure 7. Flow diagram for filter columns Seattle pilot plant.
From Kirmeyer (42)
-------
Having three columns in parallel also allowed comparisons to be made among
three different types of filter media at one time.
Mixing Intensities
Mixing intensities for each unit process were estimated using both field
data and empirical calculations. The data are tabulated and presented in
Tables 15 and 16. Detailed information on velocity gradient calculations and
measurements can be found in the EPA report written by Kirmeyer (42).
TABLE 15. VELOCITY GRADIENT DATA - SEATTLE PILOT PLANT
Unit Calculated Measured Gt at 20 gpm
G sec~l G sec"1
Rapid mix on WB-27 203 - 57,000
External rapid mixers using
1/3 HP motor and baffled
55 gallon drums
with 2 propellers
with 3 propellers
Kenics static mixer
2"
diameter*
1050
1300
1990
195,000
245,000
1,450
*A product of Kenics Corporation, Danvers, Mass.
Filter Media Tested-
Four different types of filter media were evaluated during the pilot
tests. The characteristics of each as supplied by the manufacturer are
presented in the Table 17.
Sampling and Analysis
Grab samples were gathered during the filter runs and analyses were per-
formed on these samples. pH and temperature measurements were performed at
the pilot plant location. Other analyses including aluminum, calcium, conduc-
tivity, alkalinity, tannin and lignin, color, iron, dissolved oxygen, suspended
solids and particle counts were conducted in the Seattle Water Department
laboratory according to the 14th edition of Standard Methods (43) (where
applicable). Asbestos analyses were performed by an electron microscopist at
the University of Washington. Trihalomethane analyses were performed by the
Laboratory Branch, Region X, EPA.
During normal filtration runs, turbidity information was taken directly
from in-line turbidimeters. When asbestos samples were gathered, turbidity
analysis was performed directly on a portion of each individual water sample
using the laboratory turbidimeter. Thus, correlations between turbidity and
other parameters such as asbestos and particle counts could be defined if they
existed. The on-line field turbidimeters were calibrated against the labora-
tory turbidimeter, which was calibrated using formazin standards.
39
-------
TABLE 16. MIXING INTENSITIES FOR FLOCCULATOR - SEATTLE PILOT PLANT
Flocculator Paddle
Rheostat
Percent RPM
0 0
10 1-3/4
20 3-1/2
30 5-1/2
40 8
50 10
60 12
70 14
80 15
90 18
100 19
Tip Velocity
ft/sec
0
0
0
0
0
1
1
1
1
1
0
.16
.33
.52
.75
.94
.13
.31
.41
.69
.78
Empirical
Calculations
G, sec'1 GT*
0
5
15
30
52
73
96
120
134
176
193
0
2940
8820
17640
30576
42924
56448
70560
78792
103488
113484
Transducer Output (Tq) Torque Meter Data
Without With Avg Net
Water Water Output
ft - Ib ft - Ib ft - Ib G, sec'1 GT*
0
2 -
2 -
2 -
2 -
2 -
2 -
2 -
2 -
2 -
3 -
10
10
10
10
10
10
10
10
10
10
2
4
5
6
8
9
11
14
20
20
0
- 10
- 11
- 13
- 16
- 19
- 22
- 26
- 30
- 25
- 28
0
0
1.5
3
7
7.5
9.5
12.5
16
21.5
23
0
0
30
53
98
114
140
174
204
259
276
0
0
17640
31164
57624
67032
82320
102310
119950
152290
162290
*Detention time assumed to be 9.8 minutes at 16 gpm from Table 1.
From Kirmeyer (42)
-------
TABLE 17. CHARACTERISTICS OF MEDIA TESTED - SEATTLE PILOT PLANT
Effective Uniformity Thickness Effective Uniformity Thickness
Size Coefficient inches Size Coefficient inches
(mm) (mm)
Neptune Microfloc Mixed Media*
Type MM Type CMM
Anthracite
Coal
Sand
Fine Garnet
Anthracite
Coal
Sand
Notes: MM
CMM
FC
CC
mm
1.0 - 1.1
0.42 - 0.55
0.18 - 0.32
0.92
0.40
= Mixed media
= Mixed media
= Dual media
= Dual media
= Millimeter
1.7 18 1.0 - 1.1 1.7 21
1.8 9 0.42 - 0.52 1.4 7
2.2 3 0.18 - 0.32 2.2 2
Turbitrol Dual Media^
Type FC Type CC
1.28 20 1.1 1.31 20
1.30 10 0.40 1.30 10
, sand MS -6 *Neptune Microfloc, Corvallis, Ore.
, sand MS-18 'Taulman Co., Atlanta, Georgia
with fine coal
with coarse coal
From Kirmeyer (42)
Early in the pilot studies, air bubbles that were formed in the granular
media filter would become entrained in the finished water and give false tur-
bidity readings. This situation was eliminated by installing bubble traps on
the lines that led to the recording turbidimeters.
Asbestos fiber count of the water samples was determined by the trans-
mission electron microscope method as recommended by EPA (44). Analytical
work was done by the University of Washington School of Public Health.
STUDY RESULTS
Raw Water Quality
The Tolt water supply is a high quality source of water originating from
rainfall and snowmelt runoff in the north Cascade mountains and possesses
water quality characteristics, as shown in Table 18.
41
-------
TABLE 18. RAW WATER QUALITY CHARACTERISTICS - SEATTLE
Parameter
Value
pH
Alkalinity
Hardness
Conductivity
Dissolved Oxygen
Temperature
Aluminum
Color
Tannin/Lignin
Corrosivity
Turbidity
Bacteriological Counts
Amphibole
Chrysotile
6.65 (units)
5.0 (mg/L CaO>3) .
9.0 (mg/L CaC03)
24 (micromhos)
13 (mg/L, Saturated)
2-10 ["Centigrade (°C)]
0.21 (mg/L)
18 (units)
0.25 (mg/L)
Highly Corrosive
Range: 0.10-5 ntu; Average = 0.75 ntu
Range: 1-65/100 milliliters (mL);
Average = 9/100 mL
Range <0.04 X 106 - 5.7 X 106 fibers/liter;
Average = 1.6 X 106 fibers/liter
Range 1.2 X 106 - 25.8 X 106 fibers/liter;
Average = 7.1 X 106 fibers/liter
From Kirmeyer (42)
Turbidity
Turbidity exceeds the Seattle Water Department's goal of 1.0 ntu on a
seasonal basis. It normally drops throughout the summer months to near 0.10
ntu and then begins to rise slowly with the fall precipitation. Highest tur-
bidities normally occur during January, February, March, April and December.
The turbidity has not exceeded the 5 ntu maximum contaminant level since the
National Interim Primary Drinking Water Regulations became effective in June
of 1977.
Asbestos Counts
Table 19 lists the results of both raw water amphibole and chrysotile
counts which were gathered during the study. These fibers are present natur-
ally in the streams which feed the South Fork Tolt Reservoir and have been
found in several water supplies throughout western Washington.
42
-------
TABLE 19. RAW WATER ASBESTOS COUNTS - SEATTLE PILOT PLANT
Date
of
Collection
Jan.
Feb.
Feb.
Feb.
Feb.
Mar.
Mar.
Apr.
Apr.
May
May
June
June
June
July
Sept
Oct.
Nov.
Nov.
Jan.
Jan
Feb.
June
Sept
24
2,
9,
17
23
3,
25
11
21
6,
18,
1,
9,
29
13
. 2
6,
7,
16
12
12,
14
8,
. 4
, '77
'77
'77
, '77
, '77
'77
, '77
, '77
, '77
'77
'77
'77
'77
. '77
, '77
, '77
'77
'77
, '77
, '78
'78
, '78
'78
, '78
Filter
Run
Number
3
4
5
6
11
12
21
24
29
33
44
51
53
62
70
89
93
108
111
120
120
135
161
174
Turbidity
ntu
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
3
3
1
0
0
.4
.4
.4
.3
.15
.0
.66
.60
.62
.61
.56
.54
.50
.35
.35
.35
.38
.55
.85
.30
.40
.80
.30
.36
Amphibole
fibers/liter
5
3
3
3
4
1
2
2
0
0
0
< 0
0
< 0
< 0
< 0
< 0
< 0
< 0
0
< 0
< 0
< 0
0
.70(106)
.31(106)
.06(106)
.46(106)
.33(106)
.76(106)
.18(106)
.4 (106)
.94(106)
.65(106)
.90(106)
.29(106)
Chrysotile
fibers/liter
8.89(106)
(BDL)
.70(106)(NSS)
.12(106)
.07(106)
(BDL)
(BDL)
.05(106)(BDL)
.07(106)
.14(106)
,14(106)
.19(106)
(BDL)
(BDL)
(BDL)
(NSS)
.10(106)(BDL)
.07(106)
(BDL)
.04(106)(BDL)
,07(106)
(NSS)
5
16
13
13
13
25
9
4
3
2
8
3
2
2
1
3
3
4
5
11
3
2
1
. 12 (
.39(
106)
106)
.0 (106)
.29(
.14(
.80(
.40(
106)
106)
106)
106)
.25(106)
.82(
.80(
.40(
.60(
.52(
.81(
.20(
.60(
.61(
.62(
,38(
.56(
.90(
.00(
.84(
106)
106)
106)
106)
106)
106)
106)
106)
106)
106)
106)
106)
106)
106)
106)
BDL = Below Detectable Limits
NSS = Not Statistically Significant
From Kirmeyer (42)
Amphibole fibers range in length from 0.3 to 7.5 microns and counts have
ranged from <0.04 X 106 fibers/liter up to 5.7 X 106 fibers/liter. Amphibole
counts fluctuate with the season of the year and appear to be related to raw
water turbidity. Chrysotile fibers range in length from 0.1 up to 8 microns
and counts have ranged from 1.2 X 106 up to 25.8 X 106 fibers/liter.
Filtered Water Quality
Quality Goals
Removal of all amphibole and chrysotile asbestos was the primary goal of
the research effort. To effect good fiber removal, a finished water turbidity
goal of <_ 0.10 ntu was established. Turbidities of 0.10 ntu are readily
attainable with filtration and such techniques had been effective at removing
amphibole fibers down to the detection limit at Duluth, Minnesota (12). To
preclude problems with reflocculation in the distribution system, the American
Water Works Association (AWWA) goal of <_ 0.05 mg/L for aluminum in the finished
water was also established. Because EPA had set no standard for asbestos
43
-------
levels in drinking water at the time of the Seattle study, the water quality
goals were selected by the Seattle Water Department.
Alum Coagulation
The first treatment technique tried for conditioning the water for filtra-
tion was alum coagulation. Lime was used for pH control. With an alkalinity
of only 5 mg/L as CaCC>3, the water was very sensitive to both alum and lime
addition, and numerous filter runs did not meet water quality goals because of
a slight under or over feed of lime. The problem persisted so tests were
initiated to better define the optimum pH range for destabilization. Figure 8
was developed and indicates that pH should be maintained between 6.1 and 6.7
pH units. Field data indicated that between 1 and 4 mg/L of Ca(OH)2 is re-
quired to maintain the pH in this range at a dosage of 10 mg/L of alum. The
AWWA goals of 0.05 mg/L for aluminum and 0.10 ntu for turbidity in the fin-
ished water can consistently be met in this pH range. Although these data
were collected from numerous filter runs conducted under various water quality
conditions, they indicate that pH is a very critical factor in the destabili-
zation of Tolt water.
Based on laboratory jar tests, an alum dosage of 10 mg/L was chosen as a
starting point for the pilot filter runs. This dosage, in conjunction with
granular media filtration, was very successful at removing turbidity from the
raw water under practically all conditions encountered during the testing
phase. Finished water turbidity would normally drop rapidly to j< 0.10 ntu
within a half an hour and would remain at that level until breakthrough or
terminal headless occurred. The addition of lime for pH control was critical
for consistent removal as was the use of a filter aid. (Approximately 100
filter runs were conducted at a dosage of 10 mg/L of alum.)
To determine minimum dosages for effective turbidity removal, several
filter runs were conducted at dosages ranging from 3 to 10 mg/L. The informa-
tion from these runs is summarized in Figure 9.
Based on these pilot tests, the removal of turbidity down to levels
_< 0.10 ntu is attainable using alum, lime and a filter aid in conjunction with
granular media filtration. The preferred dosage ranges and chemicals are
listed in Table 20.
TABLE 20. SEATTLE PILOT PLANT PREFERRED CHEMICAL TREATMENTS - ALUM
Chemical Dosage
Alum 7-10 mg/L
Lime [ Ca(OH)2 ] to.pH between 1-4 mg/L
6.1 and 6.7; alkalinity 4 mg/L
Nonionic or anionic filter aid 0.02 - .25 mg/L
From Kirmeyer (42)
44
-------
0.4r
0.3
g
m
or 0.2
D
0.1
TURBIDITY
, 0.10
0.08 _
\
ro
0.06 D
z
D
0.04 <
0.02
5.7
5.9
6.1
6.3 6.5 6.7 6.9
pH
7.1
From Kirmeyer (42)
Figure 8. Finished water turbidity and aluminum residuaE vs.
Seattle pilot plant.
-------
3
+<
C
CQ
CC
D
I-
CC
UJ
O
111
X
CO
i.or
0.8
0.6
0.4
0.2
8
10
ALUM DOSAGE, mg/l
Figure 9. Finished water turbidity vs. alum dosage Seattle
pilot plant
From Kirmeyer (42)
46
-------
Alum and Cationic Polymer
One of the objectives of the study was to investigate the effectiveness
of three conditioning chemicals in conjunction with granular media filtration.
To determine the optimum chemical dosages, various combinations of alum, lime,
cationic polymer and filter aids were investigated and are listed in Table 21.
TABLE 21. ALUM AND LIME DOSAGE VS. CATIONIC POLYMER DOSAGE -
SEATTLE PILOT PLANT
Alum Dosage
mg/L
2
4
8 + Lime
16 + Lime
CATFLOC T-l* Dosage
mg/L
0.2
0.2
0.2
0.2
0.4
0.4
0.4
0.4
0.8
0.8
0.8
0.8
1.6
1.6
1.6
1.6
3.2
3.2
3.2
3.2
*Calgon Corporation, Pittsburgh, Pa.
From Kirmeyer (42)
Based on these screening tests, a combination of 3-5 mg/L of alum, 2
mg/L of CATFLOC T-l and a filter aid was found to be quite effective for re-
moving turbidity down to _< 0.10 ntu. The major advantage associated with
this chemical and dosage combination is that pH adjustment with lime is not
necessary. Because much smaller amounts of alum are used, both pH and alka-
linity are only slightly affected.
Several nonionic and anionic filter aids including 1986N' CA-233*,
CA-253*, and A-23* were tested during this portion of the pilot studies. These
aids prevented rapid breakthrough of the floe particles from occurring. For ex-
ample, two filter runs indicated that although turbidity was removed to _< 0.10
ntu with a dosage of 0.1 mg/L of CA-233, breakthrough would occur several
hours before reaching terminal headloss. Increasing the dosage to 0.3 mg/L in
two other runs prevented breakthrough from occurring until terminal headloss was
reached. Table 22 shows the preferred doses for alum and polymers.
TABLE 22. SEATTLE PILOT PLANT PREFERRED CHEMICAL TREATMENTS -
ALUM PLUS CATIONIC POLYMER
Chemical
Dosage
Alum
Cationic Polymer
Filter Aid (Nonionic or Anionic Polymer)
3-5 mg/L
2 mg/L
0.1 - 0.3 mg/L
'American Cyanimid, Wayne, N.J.
Chemical Co., Midland, Mich.
From Kirmeyer (42)
47
-------
Cationic Polymer Alone
Pilot tests were run to determine the effectiveness of using a cationic
polymer without alum. The dosages of cationic polymer, CATFLOC T-l* and 573C^,
were increased slowly at set intervals during separate filter runs with the
following results.
As indicated in Tables 23 and 24 and in Figures 10 and 11, additions of
CATFLOC T-l and 573C were effective at removing turbidity down to_< 0.10 ntu
if the dosage was increased to about 3 mg/L. The advantages associated with
TABLE 23. CATFLOC T-l DOSAGES AND TURBIDITY -
SEATTLE PILOT PLANT
CATFLOC T-l Dosage (mg/L)
Finished Water Turbidity (ntu)
0.12
0.20
0.40
0.60
0.80
1.2
2.0
3.0
0.35
0.34
0.34
0.32
0.33
0.25
0.15
0.1
*Calgon Corporation, Pittsburgh, Pa.
From Kirmeyer (42)
this chemical treatment are two-fold. First, it does not affect the pH or
alkalinity of the water; thus, pH control with lime is not necessary.
Secondly, the sludge generated from cationic polymers is normally denser and
easier to dewater than an alum sludge.
Ferric Chloride
A limited amount of testing was conducted using ferric chloride at dos-
ages ranging from 3 to about 20 mg/L. Results were not very encouraging. A
dosage of 6-9 mg/L with a filter aid could reduce turbidity to about 0.2 ntu
for a short period of time. Breakthrough would normally occur rapidly and
results were inconsistent. Note, ferric chloride results were not encouraging
at Duluth either.
Evaluation of Unit Process and Treatment Train Variations
Several aspects of pretreatment and the filtration process were investi-
gated to determine their influence on both asbestos fiber removal and treat-
ment cost. Rapid mixing and flocculation were evaluated. Four different
'American Cyanamid, Wayne, N.J.
48
-------
0.4h
D
+J
c
0.3
m
DC
cc
UJ
Q
UJ
x
0.2
0.1
0
1 2
CATFLOC T-1 DOSAGE, mg/l
From Kirmeyer (42)
Figure 10. Finished water turbidity vs. Catfloc dosage -
Seattle pilot plant.
49
-------
0.20 -
0.15
to
cc
-------
TABLE 24. 573C DOSAGES AND TURBIDITY - SEATTLE PILOT PLANT
Dosage (mg/L) Finished Water Turbidity (ntu)
0.12
0.20
0.40
0.60
0.80
0.19
0.18
0.18
0.19
0.14
1.6 0.095
2.4 0.10
3.2 0.08
'American Cyanamid, Wayne, N.J. From Kirmeyer (42)
filter medias were used during the research, and a wide range of filtration
rates was employed. Data were obtained for both fixed rate and declining
rate modes of filtration.
Water Production Efficiencies
A major factor in determining the cost of a filtration plant is the sur-
face area of the filters. Filter surface area needed for production of a
given amount of water in a day's time is a function of filtration rate and
production efficiency. Three different methods, Unit Filter Run Volume (UFRV),
Percent Efficiency, and Net Water Produced per 24 Hours were used by Kirmeyer
(42). Each method involves a somewhat different approach to evaluation of
filter efficiency.
In this report, net water production was evaluated. This was defined as
the amount of usable water produced per unit surface area of a filter per 24
hours. Net water produced includes corrections for water used during backwash
and for production lost during the time of the backwash cycle. This was Net
Water Produced per 24 Hours as calculated by Kirmeyer. For this report a
correction was also made for water above 0.10 ntu produced at the beginning of
a filter run during the filter ripening period that should be filtered to
waste. The formula is:
NP =(LR X 1440) (h/H) - NBW[BW + (LR X T)]
where NP = Net water produced, gallons/ft^/24 hours
LR = Average filter loading rate, gpm/ft^
51
-------
NEW = Number of backwashes per 24 hours = 24/H
BW = Amount of backwash water used (200 gallons per ft2 per backwash,
for Seattle data analysis)
T = Time, 15 minutes down time per backwash
h = Hours of production of water with turbidity _< 0.10 ntu
H = Length of run to termination or turbidity of 0.2 ntu, hours
Because production of water with a turbidity of 0.10 ntu or less was a
goal of this study, water with higher turbidity was not included as net water
produced, even though filter run length was taken as time from start of run to
termination or 0.2 ntu turbidity in the filtered water. The calculation is
made as if the plant operated in a filter-to-waste mode when turbidity was
high at the start of a filter run.
The amount of water used per backwash was estimated at 200 gallons by
Kirmeyer (42), because the filters used in the pilot plant study were too
small to yield backwash data that would be representative of performance at a
large filtration plant. Actual backwash water used could vary with different
treatment modifications, particularly coagulant chemicals used (alum vs.
polymer), filter media sizes and filtration rate, so the limitations of this
concept should be kept in mind as it is applied to pilot plant data. In spite
of these limitations, net water produced is a useful benchmark against which
to evaluate the effects of treatment modifications.
Unit Processes
Static mixersSeveral filter runs were conducted to determine if static
mixers would effectively blend the treatment chemicals with the raw water and
if so, how many units would be needed to accomplish this task. Efficiency
data from four consecutive filter runs are listed in Table 25 and can be used
to compare results when different numbers of mixers were in use.
TABLE 25. COMPARISON OF ONE VS. THREE STATIC MIXERS -
SEATTLE PILOT PLANT
Run
No.
6A
6B
6C
6D
No. of
Static
Mixers
3
3
1
1
Net Water Produced
24 Hours
(gal/ft2/24 hours)
3785
3585
3868
3759
52
-------
Review of the filter efficiencies resulting from the use of one or three
static mixers indicates little difference in the results. Because a single
static mixer appeared to mix the treatment chemicals with the raw water as
well as or better than three mixers in series, most runs conducted during the
study utilized only one mixer.
Back mixersSeveral back mix systems were investigated including hydrau-
lic and mechanical mixers with and without the flocculator in the treatment
train. Little difference could be detected among the various back mix systems
tested. Filter runs conducted with these systems would normally be terminated
because of turbidity breakthrough well before terminal headloss was reached.
Comparison of mixing systemsBoth static and back mix systems could be
operated to produce an acceptable quality finished water. One difference that
was noted was the reason for terminating a filter run. With the static mixers,
head loss built up more quickly, and the run would be terminated because the
headloss reached 8 ft, rather than because of turbidity breakthrough. The
opposite was true with the back mix systems. Headloss would build at a slower
rate, and the run would normally be stopped because of turbidity breakthrough,
not because it reached terminal headloss. The reason for this difference in
operational characteristics was not determined.
FlocculatorTo determine if a flocculation basin should be included in
the full-scale treatment plant, filter runs 127M to 137M were conducted
with and without this unit process in the treatment train.
Review of the data in Table 26 indicates that when the flocculator was
deleted from the treatment train, water production efficiencies were consis-
tently much lower. The runs conducted at 5-6 gpm/ft^ were unable to remove
turbidity down to < 0.10 ntu and even the runs conducted at 4 gpm/ft^ experi-
enced rapid breakthrough.
When the flocculator was placed back into the treatment train, the
net water produced rose immediately to much higher levels. The turbidity
was normally held to _< 0.10 ntu until terminal headloss occurred. The
reason for the increased production with the flocculator in place was not
determined but the differences in operational capabilities were evident.
These differences were most noticeable when the raw water turbidity exceeded
1.5 ntu and water temperatures were cold, 5-6°C. The cold water conditions
may have required more contact or reaction time before filtering.
Filter Media Comparison
Removal of asbestos fibers was a prime consideration of the study, so a
comparison of the chrysotile counts in the finished water generated from the
various medias was made and results are shown in Table 27.
When the complexity of the asbestos analysis is taken into consideration,
the data indicate that all filter medias are equally effective at removing
asbestos fibers during the normal course of a filter run. With the exception
of samples collected during a turbidity spike and at the very beginning of the
filter run #174, all samples had BDL or NSS levels of chrysotile. Finished
53
-------
TABLE 26. RESULTS FROM FILTER RUNS WITH AND WITHOUT FLOCCULATION - SEATTLE PILOT PLANT
Ln
Filter
Run
No.
127M
128M
12 9M
130M
13 1M
13 2M
13 3M
13 4M
13 5M
13 6M
137M
Flocculator
Yes/No
Yes
Yes
No
No
No
No
No
Yes
Yes
Yes
Yes
Filter Loading Rate
gpm/f t2
Maximum Average
7.0
8.0
6.1
6.0
4.0
5.0
6.0
4.0
4.0
5.0
6.0
6.8
6.5
6.1
6.0
4.0
5.0
6.0
4.0
4.0
5.0
6.0
Number
of Hours
<0.10 ntu
10
16
0
0
8
5
0
25
23
16
15
Hours to
End of Run
or 0.2 ntu
13
16
-
~"
10
7
-
26
24
16
16
Net
Production
gal/ft2/day
6970
8910
0
0
3980
4200
0
5300
5260
6790
7660
Comments
Turbidity
>0.10 ntu
Turbidity
>0.10 ntu
Turbidity
>0.10 ntu
-------
TABLE 27. CHRYSOTILE RESULTS FROM VARIOUS FILTER MEDIA -
SEATTLE PILOT PLANT
Run Hour
No. Into Run
1
4
161 10
13
15
0
1
1 1/2
2
3
174
4
5
6
7
14
MM*
10 6
fibers/liter
O.Ol(NSS)
0.02(NSS)
< O.Ol(BDL)
< O.Ol(BDL)
< O.Ol(BDL)
0.07
0.1
0.36
0.03CNSS)
< O.Ol(BDL)
< O.Ol(BDL)
O.Ol(NSS)
O.Ol(NSS)
O.Ol(MSS)
< O.Ol(BDL)
Filter Column
FC* CC*
106 106
fibers/liter fibers/liter
< O.Ol(BDL) < O.Ol(BDL)
0.04(NSS) < O.Ol(BDL)
0.02(NSS) O.Ol(NSS)
0.03(NSS) < O.Ol(BDL)
< O.Ol(BDL) < O.Ol(BDL)
O.Ol(NSS)
0.04(NSS)
< 0.03(BDL)
O.Ol(NSS)
< O.Ol(BDL)
< O.Ol(BDL)
< O.Ol(BDL)
< O.Ol(BDL)
0.06(NSS)
< O.Ol(BDL)
CMM*
106
fibers/liter
-
-
-
-
-
0.09
0.02(NSS)
0.94
< O.Ol(BDL)
0.03(NSS)
O.Ol(NSS)
0.06(NSS)
0.02(NSS)
O.Ol(NSS)
O.Ol(NSS)
*Data on filter media are on p. 41 of this report.
From Kirmeyer (42)
water turbidities were also virtually the same for all of the medias being
tested.
Kirmeyer calculated water production data for a large number of filter
runs made with filters utilizing mixed media with sand MS-6, dual media with
fine coal, and dual media with coarse coal. Data are presented as net water
produced per day vs. filter loading rate in Figure 12. Differences in water
production for the three kinds of media are slight, if any exist. Water pro-
duction data in Figure 12 were not adjusted for production of filtered water
having turbidity above 0.10 ntu. In the runs used as a data base for Figure
12, filtered water turbidity was almost always j< 0.10 ntu, so any correction
55
-------
ll,000r
CO
k.
JI
CJ
\
f>J
(0
U)
Q
LJJ
O
D
Q
O
CC
0.
cc
in
h-
LLI
10,000
9,000
8,000
7,000
6,000
5,000
4,000
Legend MMS A
FC '*
456789
FILTER LOADING RATE, gpm/ft2 From Kirmeyer (42)
Figure 12. Net water produced per 24 hours vs. filter loading rate
56
-------
factor for the net production would be small. Filter performance was affected
more by pretreatment conditions than by media types in experiments with Tolt
Reservoir water.
Summary of Treatment Results
Fiber count data from the Seattle study are compiled in Appendix B of
this report. A summary of the appendix data is presented in Table 28. It
shows that chrysotile removal exceeded 99.0 percent in 50 of 125 samples.
Removal was greater than 95.0 percent for 100 of 125 samples. Amphibole re-
moval exceeded 99.0 percent in 6 of 18 samples and 95.0 percent in 15 of 18
samples. Chrysotile fiber counts were below detectable limits (BDL) or not
statistically significant (NSS) for 46 of 125 filtered water samples.
Amphibole fiber counts were BDL or NSS for 15 of 18 filtered water samples.
Most of the samples with filtered turbidity above 0.10 ntu had fiber removals
of 95.0 percent or lower.
The Seattle study confirmed the Duluth pilot plant research results. In
addition, the work at Seattle included runs to provide simultaneous comparison
of different kinds of filter medias and extensive asbestos sampling throughout
certain filter runs to give a better understanding of fiber removal by granu-
lar media filters.
57
-------
TABLE 28. SUMMARY OF FIBER COUNT DATA FOR SEATTLE STUDY
00
Filtered Number
Water of
Fiber samples
Removal in this
percent category
> 99.0
> 95.0
99.0
> 90.0
95.0
_< 90.0
> 99.0
) 95.0
99.0
> 90.9
95.0
Chrysotile Data
50
to 50
to 13
12
Amphibole Data
6
to 9
to 1
Percent of Number
samples of samples
in this with
category filtered
turbidity
< 0.10 ntu
Number Number Percent
of samples of samples with
with with BDL
filtered BDL fiber
turbidity fiber count
> 0.10 ntu count
Base: 125 filtered water samples
40%
40%
10%
10%
Base: 18
33%
50%
6%
raw water
48
43
7
3
filtered water
raw water
6
9
0
Number Percent
of samples with
with NSS
NSS fiber
fiber count
count
paired with statistically significant
chrysotile counts
2
7
6
9
samples
amphibole
0
0
1
11
3
0
0
paired with
counts
3
6
0
22%
6%
0%
0%
statistically
50%
67%
0%
15
17
0
0
significant
3
3
0
30%
34%
0%
0%
50%
33%
0%
< 90.0 2 11% 0 2 0 0% 0 0%
-------
SECTION 7
SAN FRANCISCO BAY AREA
During the course of preparing a doctoral dissertation entitled, "Asbes-
tos in Drinking Water and Cancer Incidence," Kanarek incidentally obtained
some data on asbestos fiber removal at water filtration plants (19). Col-
lecting information about treatment was not the main purpose of Kanarek1s work,
so only a portion of the treatment plant data he obtained could be used in
this report. Data from nine treatment plants owned and operated by seven
utilities were compiled. The water sources, as well as the raw water asbestos
fiber counts, varied greatly.
Kanarek's data are most useful for the present effort when both raw and
filtered water fiber counts are given. Removal percentages can not be calcu-
lated without raw water data, nor can adequacy of treatment be discussed with
confidence if treated water fiber counts are very low but raw water counts are
unknown, and possibly also very low. For these reasons caution must be used
in interpreting the data presented by Kanarek.
Treatment performance data to supplement Kanarek's asbestos data were
furnished by the Sanitary Engineering Section of the Department of Health Ser-
vices, Health and Welfare Agency, State of California. The data, although
incomplete, show that treatment practice varies in the Bay area. Some plants
receive raw water of excellent quality, as measured by usual parameters such
as turbidity. Some of these are direct filtration plants. At times some of
the plants filter this high quality raw water with little or no coagulant
chemical added to condition the water. Other plants in the area, such as
those in Contra Costa County, treat turbid waters that are subject to waste-
water contamination by upstream outfalls. These plants employ complete treat-
ment, and they tend to strive for the best possible filtered water quality.
TREATMENT PLANTS IN CONTRA COSTA COUNTY
Three of the treatment plants for which some data are available are loca-
ted in Contra Costa County. According to Kanarek (19), filtration plants for
the City of Antioch, City of Pittsburg, and Contra Costa County Water District
(Bollman Plant) draw raw water from the Contra Costa Canal, the Sacramento
River, or the San Joaquin River (Fig. 13).
Design and Operating Information
Bollman Plant
The Bollman plant is the largest of the plants in Contra Costa County for
59
-------
Sacramento-San Joaquin Delta System
Sacramento River
Municipal wastewater and
asbestos industry discharge
Contra Costa Canal
Rock Slough Intake
San Joaquin River
Asbestos industry
discharge
Mallard Reservoir
Pittsburg Treatment Plant
Bollman Treatment Plant
Martinez Treatment Plant
Figure 13. Water sources for water treatment plants.
Contra Costa County
60
-------
which data are available. Stone indicated that this plant could pump water
from the Sacramento River when flows were great enough to overcome tidal
effects from San Francisco Bay (45). According to Harris the turbidity of the
raw water at the Bollman Plant was expected to range from 25 to 300 ntu (46).
He wrote that the drainage of the Sacramento and San Joaquin River systems con-
tributed' considerable amounts of industrial, municipal, and industrial wastes
to the source water.
Because of the raw water quality, the Bollman Plant was designed to pro-
vide very thorough treatment to the water. Treatment includes coagulation,
flocculation, sedimentation and filtration. Plant design criteria as given by
Stone are shown in Table 29. The Bollman Plant was designed so that alum,
lime, caustic soda, and polymer could be fed to adjust pH and condition the
water for filtration.
Filter media were evaluated in pilot filter studies. Media selection is
given in Table 30. The combined media depth after backwashing was 31 inches.
Stone noted that the top of the filter media was 7.1 feet below the surface of
the water when the filter was in operation. This design was used so that a
positive head could be provided in the filter media under all operating con-
ditions.
Antioch and Pittsburg Plants
Engineering information about the Antioch and Pittsburg plants was pro-
vided by the Sanitary Engineering Section of the Department of Health Services,
State of California. The Antioch plant uses upflow clarifiers with sludge re-
circulation. The floe zone provides about 15 minutes detention time. Deten-
tion time in the settling zone is about 50 minutes. The surface loading for
the upflow clarifier is about 2 gpm/sf. Alum is used as the primary coagulant,
and polyelectrolytes are used as both coagulant and filter aids. The filters
consist of 20 inches of anthracite on top of 10 inches of sand. During 1976-
77 the filtration rate ranged from 1.5 to 3 gpm/sf. The treatment plant at
Pittsburg is of conventional design with rapid mix, flocculation, about 4
hours of detention time in the settling basins, and filtration through dual
media at 2 to 2.5 gpm/sf. Specifications for the media are:
i.
ii.
10 inches of
20 inches of
Sieve Size
U.S. #16
U.S. #20
U.S. #30
U.S. #40
U.S. #50
anthracite 1.5-1.6 mm, U.C.<1.70
crystal Monterey sand, graded as follows:
% Passing Sieve
96-99%
68-89%
19-58%
2-7%
0-1%
61
-------
TABLE 29. DESIGN AND HYDRAULIC INFORMATION FOR BOLLMAN PLANT,
CONTRA COSTA COUNTY WATER DISTRICT
Plant capacity, mgd 80
Rapid mix, minutes 0.9
Flocculation
high energy, minutes 11
low energy, minutes 7.5
Sedimentation, minutes 60
Filters
Number (with 2 bays each) 4
Square feet per filter 1447
Flow per filter, mgd 20
Maximum Filtration rate, gpm/sf 9.6
TABLE 30. BOLLMAN PLANT FILTER MEDIA
Specific Depth
Gravity inches
Top Layer Anthracite 1.48 16
Intermediate Layer Anthracite 1.58 8
Lower Layer Silica Sand 2.6 10
Effective Size
mm
1.58
0.89
0.3
Water Quality
Raw Water
Asbestos fiber counts and turbidity data for the Contra Costa County raw
water samples are given in Table 31. The fiber data contain only three statis-
tically significant chrysotile values, and this at first appears to make an
assessment of treatment efficacy difficult or impossible. Careful considera-
tion of the data, however, reveals a tendency for raw water samples to contain
asbestos fibers in the range of 106 f/L to 107 f/L. The three samples that
did contain statistically significant fiber counts had 8, 4, and 17 X 106 f/L.
Two samples of raw water taken at the CCCWD had fiber counts that were
not statistically significant, but because of the debris present in the raw
water and the small amount of water that could be filtered for analysis, the
finding of a single fiber represented 2.5 or 5 X 106 f/L. These sample re-
sults, although not statistically significant, suggest that the raw water
fiber count was in the millions. Of the 8/13/76 samples from the canal, one
62
-------
TABLE 31. RAW WATER DATA FOR CONTRA COSTA CANAL AND OTHER SURFACE WATERS
IN CONTRA COSTA COUNTY
Date
12/10/74
10/13/75&
10/14/75
5/8/76
5/8/76
8/13/76
8/13/76
8/12/76&
8/13/76
4/18/77
4/18/77
4/18/77
4/18/77
5/11/77
1/9/78
4/4/78
PH
Location
Mallard Reservoir 7.7
CCCWD
North Inlet 7.7
Mallard Reservoir
Pittsburgh raw -
Antioch raw
Contra Costa 7.7
Canal @ Clyde
Contra Costa 7.5
Canal @ Pittsburg
Contra Costa 7.6
Canal @ Antioch
Clyde, CCCWD 7.7
intake on Canal
Contra Costa 7.4
Canal @ Pittsburg
Contra Costa 8.2
Canal @ Antioch
Clyde - CCCWD 7.9
settled 8 days
Contra Costa 7.7
Canal @ Clyde
Mallard Reservoir 7.6
from canal @ Clyde
settled 3 days
Mallard Reservoir 7.6
from Mallard Slough
Turbidity
ntu
23
40
42
-
32
28
18
25-19
21
25
3
29
22
38
Chrysotile
106 f/L
0.03 NSS
(0.03)
0.6 NSS
(0.2)
21
BDL(4.3)
present,
impossible
to count
present,
impossible
to count
8
10 NSS
(2.5)
-
-
0.6 NSS
(0.2)
15 NSS
(5)
4
17
Amphibole
106 f/L
BDL(0.03)
0.2 NSS
(0.2)
BDL(4.3)
BDL(4.3)
present
impossible
to count
present
impossible
to count
0.8 NSS
(0.8)
BDL
(2.5)
-
-
BDL
(0.2)
BDL
(5)
BDL
(0.5)
3.3 NSS
(1.7)
Detection limit in parenthesis
CCCWD = Contra Costa County Water District
63
-------
was counted and had 8 X 10° f/L. Fibers were present in two other samples
taken that day, but determining the number of fibers per liter was not pos-
sible. Nevertheless, on the basis of the Antioch sample, suggesting that the
other two 8/13/76 samples contained asbestos fibers in the range of. 10^ to lO''
f/L seems reasonable.
Filtered Water
Data for turbidity and fiber counts of raw and filtered water samples
from Antioch, Pittsburg, and CCCWD are shown in Tables 32 to 34. The Bollman
plant consistently produces filtered water with a turbidity of less than 0.10
ntu. Turbidity reduction was 99.7 percent or higher. Chrysotile fiber count
reduction exceeded 99 percent twice, and on two other instances it probably
was 98 percent or more. The chrysotile count in the filtered water ranged
from below 0.02 to 0.7 X 106 f/L. Filtered water turbidities at the Antioch
and Pittsburg plants are slightly higher than at the Bollman plant, and fiber
count reduction is slightly lower. Fiber counts in the filtered water ranged
from 0.05 to 0.8 X 10^ f/L, values similar to those reported at the Bollman
plant.
SAN FRANCISCO
Both water quality and coagulation practice are more variable at the San
Andreas plant. This plant is a conventional plant, employing rapid mixing,
coagulation, sedimentation and filtration. During periods of low turbidity, a
low alum dose is used to destabilize particulates and form tiny floe particles
that do not settle but are removed in the filtration process. With respect to
the water quality characteristics in the existing EPA Drinking Water Regula-
tions, raw water quality at this plant is excellent. Turbidity before filtra-
tion sometimes meets the 1 ntu limit in the Regulations. Before the Kanarek
study was made, coagulant chemical feed was often shut off if the water met
the 1 ntu limit before filtration. Since Kanarek's findings were made known,
treatment practice calls for the use of a coagulant chemical whenever water is
filtered, regardless of raw water turbidity, in order to achieve more effec-
tive fiber removal.
As the treatment information indicates in Table 35, the plant was some-
times operated without any coagulant chemicals. Because of the characteris-
tics of the raw water, the doses of alum and polymer, when used, are low.
During the 1975 to early 1977 period of operation, asbestos fiber counts were
higher by a factor of 2 to 4 than they have been since May, 1977.
MARIN COUNTY
Kanarek's study included data on the Marin Municipal Water District's
San Geronamo and Bon Tempe plants. According to Kanarek, these are conven-
tional plants, employing flocculation, sedimentation, and filtration (19).
Sources of raw water are shown in Figure 14. Water for the District comes
from local runoff from the Coast Range. It is collected in reservoirs in the
Mt. Tamalpais region and treated at the two filtration plants. Raw water data
are limited to two samples taken at the San Geronamo plant (see Table 36).
Samples of treated water show that both turbidity and fiber counts are lower
at the Bon Tempe plant. Filtered water turbidities as high as 1.0 ntu were
64
-------
TABLE 32. TREATMENT DATA FOR ANTIOCH PLANT
Coagulation Raw
Date
5/8/76
8/12/76
8/13/76
4/18/77
Chemical
Used
Alum
Polymer
Caustic*
Alum
Polymer
Lime*
Alum
Polymer
Lime*
mg/L pH Turbid-
ity
(ntu)
40 - 25
0.0061
41 7.6 18
0.0077
45
0.0058 8.2 25
Water Data Finished Water Data % Reduced
Chryso- Amphi- pH Turbid-
tile bole ity
(106 f/L) (ntu)
BDL BDL - 0.14
(4.3) (4.3)
7.4 0.21
8 0.8 NSS
(0.8)
- -
8.0 0.26
Chryso- Amphi- Turbid-
tile bole ity
(106 f/L)
0.13 BDL 99.4
(0.02)
98.8
0.26 BDL
(0.02)
0.13 BDL 99
(0.025)
Chryso-
tile
_
96.7
*Used to adjust pH
Detection limit shown in parenthesis
-------
TABLE 33. TREATMENT DATA FOR BOLLMAN PLANT
Date
Coagulation
Chemical mg/L
Used
Raw Water Data
Finished Water Data
% Reduced
pH Turbid- Chryso- Amphi- pH Turbid- Chryso- Amphi- Turbid- Chryso-
ity tile bole ity tile bole ity tile
(ntu) (106 f/L) (ntu) (106 f/L)
12/10/74 Alum 41
Polymer 0.02 7.7 23
Lime & Caustic*
10/13/75
10/14/75 Alum 41.5
Polymert 0.017 7.7 40
Lime & Caustic*
0.03 NSS BDL BDL
(0.03) S.I 0.04 (0.005)
0.6 NSS 0.2 NSS
(0.2)
0.13
8.06 0.04
BDL
BDL
(0.005)
99.8
99.9
Alum 45.6
8/13/76 Polymert 0.019 7.7 32
Lime & Caustic*
Alum 47.8
4/18/77 Polymert 0.022 7.7 25-19
Lime & caustic*
(settled 8 days) 3
5/11/77 Alum 49.6
Polymert 0.026 7.7 29
Lime & Caustic*
Alum 54.1
8/5/77 Polymert 0.032 7.7 35
Lime & Caustic*
Present
can't count
8.2 0.05 0.16
10 NSS BDL 8.15 0.055 0.025
(2.5)
0.6 NSS BDL
(0.2)
NSS
BDL#
0.01 99.8+
NSS
§ BDL§ 99.7+ could
(0.025) be 99.75%
BDL#
(0.025) (0.025)
15 NSS BDL 8.11 0.05 0.18
(5)
!.l 0.053 0.7
BDL 99.8+ could
(0.025) be 98.8%
BDL 99.8+
(0.05)
(continued)
-------
TABLE 33. (continued)
Date
Coagulation
Chemical mg/L
Used
Raw Water Data
pH Turbid- Chryso- Amphi-
ity tile bole
(ntu) (106 f/L)
Finished Water Data
% Reduced
pH Turbid- Chryso- Amphi- Turbid- Chryso-
ity tile bole ity tile
(ntu) (106 f/L)
Alum 49.8
1/9/78 Polymer1 0.032 7.6 22
Lime & Caustic*
Alum 46.8
4/4/78 Polymer1" 0.032 7.63 38
Lime & Caustic*
BDL 8.45 0.05 BDL BDL 99.7+ 99.37+
(0.5) (0.025) (0.025)
17
3.3 8.3 0.061 0.15
NSS(1.7)
0.025 99.8+ 99.1
NSS
*Used to adjust pH
'Generally nonionic polymer used as a filter aid
Detection limit shown in parenthesis
§After filtration
*After clearwell
-------
TABLE 34. TREATMENT DATA FOR PITTSBURG PLANT
CD
Date
5/8/76
8/13/76
4/17/77&
4/18/77
Coagulation
Chemical pH
Used
Data 8.3
Unavailable
Alum 7 . 5
Lime
Alum 7 . 4
Lime
Raw Water Data
Turbid-
ity
(ntu)
42
28
21
Chryso-
tile
(106
21
Present
-
Amphi-
bole
f/L)
BDL
(4.3)
Present
-
Finished Water
pH Turbid-
ity
(ntu)
8.0 0.28-0.
7.5 0.27
7.4 0.10
Chryso-
tile
(106
38 BDL
(0.26)
0.76
0.05
NSS
Data
Amphi-
bole
f/L)
BDL
(0.26)
0.04
NSS
BDL
(0.025)
% Reduced
Turbid-
ity
99.2
99
99.5+
Chryso-
tile
98.7+
-
Detection limit shown in parenthesis
-------
TABLE 35. TREATMENT DATA FOR SAN ANDREAS PLANT
Date
Coagulation
Chemical mg/L
Used
Raw Water Data
pH Turbid- Chryso- Amphi-
ity tile bole
(ntu) (106 f/L)
Finished Water Data % Reduced
pH Turbid- Chryso- Amphi- Turbid- Chryso-
ity tile bole ity tile
(ntu) (106 f/L)
Alum 2.1
5/5/75 Magnafloc1 0.23 7.6
Caustic*
12/14/76 None
7.6
1.1 3.5 BDL 8.6 0.12 1.4 BDL 89 60
NSS
0.64 3.7 BDL 8.7 0.32 3.4§ 0.08§ 50 8
estimate (0.08) 1.4# 0.32# 62
Alum 0.75
2/17/77 Catfloc-T@ 0.69 7.5
Caustic*
Alum 2.0
5/26/77 Catfloc-T 0.75 7.5
Alum 3.0
8/17/77 Catfloc-T 0.75 7.6
Alum 2.1
1/11/78 Catfloc-T 0.75 7.5
Alum 12.5
4/12/78 Polymer none 7.6
Caustic*
0.75
1.2
0.9 to
0.76
1.4
2.2
!.6 0.15 1.9 BDL 80
(1.5 EPA) (0.25)
8.8
0.12
0.12
0.11
8.6 0.20
7.9 0.14
1.5 BDL 8.7 0.22
(0.13)
0.4 BDL(0.025) 90
0.8 0.05 NSS 90
0.32 BDL(0.025) 91
78 to
0.93 0.05 NSS 74
(0.025)
0.7 BDL 90
(0.025)
0.32 BDL 90
(0.064)
79
'American Cyanamid, Wayne, N.J.
*Used for pH adjustment
@Calgon Corporation, Pittsburgh, PA
after filtration
after clear well
-------
Bon Tempo Lake
C
Bon Tempe Treatment Plant
J>
Nicasio Reservoir
c
San Geronimo Treatment Plant
Extreme Dry
Season Only
Ranney Well, Chlorination
I
Russian River
Figure 14. Water sources for Mann Municipal Water District plants.
70
-------
TABLE 36. TREATMENT DATA FOR MARIN MUNICIPAL WATER DISTRICT AND
NORTH MARIN WATER DISTRICT
Date
Coagulation
Chemical mg/L pH
Used
Raw Water Data
Turbid- Chryso- Amphi-
ity tile bole
(ntu) (106 f/L)
Finished Water Data
% Reduced
pH Turbid- Chryso- Amphi- Turbid- Chryso-
ity tile bole ity tile
(ntu) (106 f/L)
12/2-3/74 - -
3/6/75 Alum 23 7.7 8-6
L ime 12
3/7/75 Alum
Lime
25 7.2-7.1
17
12/13/76 Alum 7.8
Polymer 0.6
Lime 10
1/10/78 Alum 51
Lime 0
San Geronomo Plant, Marin Municipal Water District
2/25/76 Alum 9 7.2 8
Polymer 1.0
Lime 8
11/12/76 Alum 14.3 7.6 19-2.1
Polymer 1.6
Lime 0
2.5
8/17/77 Alum 6.1 6.6 14
Lime 0
6.8 46
Detection limit in parenthesis
11
0.3
0.25
5.6 1.0-0.82
}.6 1.0-0.57 2
8.6
7.1
7.9
(continued)
0.55 0.14
83-90
80-89
93
0.16 0.04 0.04 92-99
NSS NSS
0.41 BDL BDL
(0.02) (0.02)
0.29 0.7
0.1
NSS
1.2 0.64-0.28 12
84
98
BDL 99
(0.5)
82
53
44
-------
TABLE 36. (continued)
Date
Coagulation
Chemical mg/L pH
Used
Raw Water Data Finished Water Data % Reduced
Turbid- Chryso- Amphi- pH Turbid- Chryso- Amphi- Turbid- Chryso-
ity tile bole ity tile bole ity tile
(ntu) (106 f/L) (ntu) (106 f/L)
Bon Tempe Plant, Marin Municipal Water District
12/13/76
8/10/77
1/14/78
Alum
Lime
Alum
Lime
data
22.6 7.7
14
14.1 8.0
13
7.3
unavailable
1.0 - - 8.4 0.25 0.2
NSS
0.70 - - 8.8 0.17 BDL
(0.1)
23 - - 7.4 0.12 BDL
(0.1)
BDL 75
(0.05)
BDL 76
(0.1)
BDL 99+
(0.1)
Stafford Plant, North Marin Water District
3/15/78
Alum
Lime
Polymer
60 7.3
24
0.03
65 5 NSS BDL 8.4 <0.05 BDL
(5) (5) (0.025)
BDL 99.9+
(0.025)
Detection limit in parenthesis
-------
reported at San Geronomo, vs. 0.25 ntu at Bon Tempe. The highest filtered
water fiber count at San Geronomo was 12 X 106 f/L, but it was only 0.2 X 106
f/L (NSS) at Bon Tempe.
The Stafford treatment plant of the North Marin Water District was sampled
one time. Both the finished water fiber count and filtered water turbidity
were quite low (Table 36). Data suggest that the raw water may contain asbes-
tos fibers, this would be expected at Stafford.
The San Francisco Bay area has many water treatment plants, with a variety
of types, including direct filtration, upflow clarification, and conventional
flocculation-sedimentation-filtration plants. Although additional data would
have been very useful for interpreting the fiber removal results, the availab"
information is very interesting and shows, in general, that water filtration
plants can remove asbestos fibers under proper operating conditions.
73
-------
SECTION 8
PHILADELPHIA
The City of Philadelphia has three water treatment plantss Torresdale on
the Delaware Estuary; and Belmont and Queen Lane on the Schuylkill River.
About half of Philadelphia's drinking water is supplied by the Torresdale
Plant and the other half is supplied by the Queen Lane and Belmont Plants.
Generally the portion of Philadelphia east of Broad Street receives treated
Delaware Estuary water while the section west of Broad Street receives treated
Schuylkill River water.
DESIGN AND OPERATING INFORMATION
The three treatment plants are typical of plants east of the Rocky Moun-
tains that have been designed for turbidity removal. All three treatment
facilities are coagulation plants of the rapid sand filter type, with auto-
matic and semi-automatic controls.
Torresdale Plant
The Torresdale Plant receives raw water to a pre-sedimentation basin via
a difference in pre-sedimentation basin and estuary water levels. Raw water
from the pre-sedimentation basin is pumped by eight 60-MGD pumps to eight
separate rapid mix chambers. Coagulant addition takes place in these chambers.
Beyond the rapid mix chambers the Torresdale Plant is operated as two parallel
facilities, a North and South section. Each section contains two 2.3 million
gallon flocculation basins, two 10 million gallon sedimentation basins and 47
rapid sand filters. Finished water is stored in two clear wells with a total
storage of 194 million gallons. The Torresdale Plant has a rated capacity of
282 MGD and a hydraulic capacity of 423 MGD.
Chemical treatment at the Torresdale Plant includes marginal chlorination
at the presedimentation basin, breakpoint chlorination during rapid mix, and
post addition of ammonia to form chloramines. Lime is added for pH stabiliza-
tion to facilitate coagulation with ferric chloride. Post chemical treatment
includes addition of fluoride. Occasionally powdered carbon and chlorine diox-
ide are used for taste and odor control.
Queen Lane Plant
The Queen Lane Plant receives raw water via four 40 MGD and two 20 MGD
pumps located on the Schuylkill River just below the confluence of the
Wissohickon Creek. The intake pumps are protected by a floating boom and
74
-------
vertical bar screens. The raw water is delivered to a 177 million gallon raw
water storage reservoir. Raw water from the storage reservoir flows by gravity
into the pretreatment building which contains four rapid mix chambers. The
Queen Lane Plant is divided into a north and south section. Each section con-
tains two 1.0 million gallon flocculation basins and two 4.3 million gallon
sedimentation tanks stacked to accomplish primary settling in the upper level
and secondary settling in the lower level. The north section contains 21 rapid
sand filters and the south section contains 19 rapid sand filters, with each
filter rated at 3 MGD. Filter water flows to north and south bi-level clear
wells, with a total combined storage of 83 million gallons. The Queen Lane
Plant has a rated capacity of 120 MGD and a hydraulic capacity of 160 MGD.
Chemical treatment at Queen Lane includes breakpoint prechlorination at
the raw water storage reservoir. Ammonia is added after filtration to form
chloramines. Coagulation is achieved through the addition of ferric chloride
and lime. Post chemical treatment also includes the addition of fluoride and
zinc phosphate for corrosion control. Occasionally powdered carbon is used
for taste and odor control.
Belmont Plant
The Belmont Plant receives raw water through submerged intakes on the
Schuylkill river. Each intake is equipped with primary and secondary vertical
bar screens. Water is pumped by two 40 MGD and two 20 MGD pumps to two raw
water reservoirs operated in series with a total capacity of 72 million gal-
lons. Raw water flows by gravity to rapid mixing facilities each containing
three chambers. The Belmont Plant is divided into a north and south section.
The north section contains two 0.55 million gallon flocculation basins followed
by two 2.36 million gallon sedimentation basins, and 12 rapid sand filters
rated at 3 MGD each. The south section contains two 0.66 million gallon floc-
culation basins, followed by two 3.24 million gallon sedimentation basins,
and 14 rapid sand filters. Filtered water flows into a 1.8 million gallon
clear well and then into a 21.7 million gallon finished water reservoir. The
Belmont Plant has a rated capacity of 78 MGD and a hydraulic capacity of 108
MGD.
Chemical treatment at Belmont includes marginal pre-chlorination followed
by breakpoint chlorination. Ammonia is added after filtration to form chlora-
mines. Coagulation is achieved through the addition of alum and lime. Post-
chemical treatment includes the addition of fluoride, and zinc phosphate for
corrosion control. Occasionally powdered carbon and chlorine dioxide are used
for taste and odor control.
Performance for all three plants is summarized in Table 37.
WATER QUALITY DATA
A cooperative study between the Philadelphia Water Department and EPA was
undertaken after a limited sampling program had indicated that raw and fil-
tered waters in Philadelphia contained asbestos. Two phases of this study have
been completed, and a third is under way.
75
-------
TABLE 37. DESIGN AND OPERATING INFORMATION FOR PHILADELPHIA PLANTS, 1977-1978
Torresdale
Queen Lane
Belmont
Rated Capacity
Peak MGD 423
Average MGD 282
Daily Output
Peak MGD 306
Average MGD 224
Flocculation Time at Average
Rated Capacity, minutes 47
Sedimentation Time at Average
Rated Capacity, minutes 204
Filtration Rate at Average
Rated Capacity, gpm/sf
Coagulant
Type
Dosage (avg) mg/L
Typical pH(Finished)
Turbidity
River
Ferric Chloride
13.3
8.5
160 108
120 78
159 90
110 73
44 North section
48 45 South section
206 188 North section
222 South section
Ferric Chloride Alum
7.3 15.5
6.9 7.2
Average, ntu 12
Range, ntu 3-80
Finished
14
2-230
19
2-390
Average, ntu
Range, ntu
0.24
0.1-1.4
0.10
0.01-1.3
0.44
0.2-1.6
76
-------
During the first phases of sampling and analysis, monitoring of raw and
filtered water quality during normal plant operation was emphasized. Seven
day composite samples of raw and filtered waters were collected at all plants.
In addition, so that effects of storms and high runoff could be evaluated, some
twenty-four hour composite samples were taken during periods of rising river
flow and falling river flow.
Effects of Storms
Because early sampling activities had produced conflicting data on asbes-
tos in raw and finished waters, effects of storm flow were studied in order to
learn if asbestos concentrations in the rivers were influenced by scouring and
deposition during flow changes. In December, 1977 two storms occurred in the
area, and the sampling was carried out.
On December 14, 0.9 inches of rain fell, according to the U.S. Weather
Station at Philadelphia International Airport. The second storm occurred on
December 21, when 1.0 inches of rain fell at the airport. Figure 15 shows the
variation of Delaware Estuary flow as measured at Trenton, New Jersey, and raw
water turbidity at the Torresdale Plant during the time period that included
both storms. Figure 16 shows similar data for the Schuylkill River at the
Belmont Plant. Substantial increases in flow occurred at both plants after
the December 14 precipitation fell at Philadelphia. The variations in flow on
December 15-20 probably were the result of precipitation in the watersheds in
areas upstream from Philadelphia. The trends that occurred during sampling
were rising flow and turbidity followed by falling flow and turbidity. Results
of the storm sampling program are given in Table 38.
Asbestos fiber counts in the rivers are higher at the beginning of a
storm, when flow is rising and sediment previously deposited on the river bed
is picked up and carried by the higher flow.
Treatment Plant Performance
In plant performance evalution studies at Philadelphia plants, 7-day
composite samples have generally been collected. Sampling for filtered water
generally has been delayed by 24 hours to allow for time of flow through the
plants. Results are shown in Tables 39-41. The results show that conventional
water treatment plants with coagulation, flocculation, sedimentation and fil-
tration can remove a high percentage of the asbestos fibers found in the raw
water. Substantial removal by sedimentation was observed when raw, settled,
and filtered water samples were analyzed. Sedimentation at Torresdale
reduced the fiber count from 15.8 to 0.7 X 106 f/L. At Belmont, sedimenta-
tion reduced the fiber count from 10 to 1.7 X 106 f/L, but the filtered
water contained 1.35 X 10^ f/L. These samples were taken as 24-hour compos-
ite samples during a period when 7-day composite samples were taken. The
7-day composite for the Belmont plant had 8.6 X 10^ f/L in the raw water
and 0.28 X 106 f/L (NSS) in the filtered water. The significance of the
1.35 X 106 f/L filtered water sample is uncertain because no other filtered
waters had such a high fiber count. Additional composite sampling of raw,
settled, and filtered water is needed.
77
-------
00
5O
)
-a
c
(0
w
^ 40
o
,£
*'
d
o
o
UJ
30
\
UJ
UJ
L1-
O
CO
-^1
0 20
10
^ Delaware estuary flow at Trenton
December 13-26
O Turbidity at Torresdale Intake
-
o
c
*'
fo ~fy ^c y
13 1
J
0 £
0 £
0 ***
* 0
*
o
*
* *
1
4 15
* n I c r »
^ rilot *
COMPOSITE
*
*** *
** * >
* *%
. **
3JC *r*
O
0
0 °°°
O o
0
0 0
i i i
16 17 18
j FALL *
COMPOSITE
STORM 1
December 14-26
:*
0* A^
0 *
^ O
O A O
* O
1
19 20
TIME, days
*&
^
**
0 *
*
*
*
* *
*
*
* 0 *J
o
o
c O
o o o
c
0
I I
21 22 23
rj I <* | p.
COMPOSITE
*
*
* *'
** *
00 * *
*
o o
0
0 0
o
I
24 25
, FALL *
COMPOSITE
STORM 11
-
%
**i
o
00
26
50
40
£
c
30 .>
9
20 m
CC
I-
10
0
Figure 1 5. Flow and turbidity during Torresdale plant storm sampling.
-------
40
w
c
ra
W
3
O
5 30
G
O
O
LLJ
V)
\ 20
r-
UJ
LL,
O
SO
o
10
_
-
-
°o°oc
]
jjl jV Iji ->& jfi jft X3
i
13 14
C
0 <
o
^
£
* £
* 0
OOOQO
I
15
RISE »-
o
° 0
0 °
0 0. °
0 0 0
0 0°°00
*..
* **»«» ,<
1 1
. 16EA1I17
;OMPOSITE COMPOSITE
* STOHM 1
o Schuylkill River Turbidity
^
*
*
*
o
^f
*
* *
*
*
IV
0 %
^
0°
1 1
18 19 20
TIME, days
at Belmont
12/14/77-12/25/77
$ Schuylkill River Flow
at Philadelphia
12/13/77 12/26/77
00
oo
0 O
Jj o
* c
*
*
° 0 X
U *
oo o
* 0
0 0
* °
I I I
21 22
-, RISE ^
COMPOSITE
oo
0
006
oo
* * ° *
* o 0 * '
$&& Q *^ e\ n
O
I I 1
23 24 25
« FALL »
COMPOSITE
* STOR vi 11
>
Xt
**_
-
100
90
80 3
c
70 .-
^
60 t
Q
50 m
DC
40 D
I
30
20
10
0
26
Figure 16. Flow and turbidity during Belmont plant storm sampling.
-------
TABLE 38. STORM FLOW SAMPLING RESULTS AT PHILADELPHIA
River/
Fin.
River
Fin.
River
Fin.
Torresdale Belmont
Storm Rise Storm Fall Storm Rise Storm Fall
Storm Asbestos Asbestos Asbestos Asbestos
No. Date Type 106 f/L 106 f/L 106 f/L 106 f/L
1 12/14/77 C
A
1 12/14/77 C
A
2 12/21/77 C
A
2 12/21/77 C
A
2.9 NSS(0.3) 7.6
BDL(0.3) BDL(0.3) BDL(0.3)
BDL(O.Ol)
BDL(O.Ol)
6.1 4.7 14.6
BDL(0.3) BDL(0.3) BDL(0.7)
0.1 - NSS(0.02)
BDL(O.Ol) - BDL(0.02)
5.4
BDL
(0.3)
-
1.7
BDL
(0.3)
-
C: Chrysotile
A: Amphibole
NSS: Not Statistically Significant
BDL: Below Detectable Limit
Detection Limit in parenthesis
80
-------
TABLE 39. WATER QUALITY DATA - TORRESDALE PLANT
oo
Raw Water
Date Turbidity Chrysotile
fibers
Avg. Range
ntu ntu 106 f/L
12/28/77-
1/3/78 8 7-12 3.6
12/21/77 29 18-50 6.1
Storm Rise
2/22-28/78 7 5-8 1.2 NSS
11/13-19/78 6 4-9 2.2
1/8-9/79 34 27-40 15.8
settled water to sand filter
sand filter effluent
settled water to GAC filter
GAG filter effluent
Filtered Water Percent
Turbidity Chrysotile Turbidity
fibers
Avg. Range
ntu ntu 106 f/L
0.13 0.11-0.17 0.02 NSS 98.3
(0.02)
0.19 0.16-0.23 0.1 98.9
0.13 0.11-0.14 0.03 NSS 98.1
(0.01)
0.27 0.23-0.30 0.28 NSS 95.5
(0.14)
See data below
3.7 3.0-4.9 0.7 NSS 89.1
(0.7)
0.38 0.34-0.40 (0.14) BDL 99.1
3.7 3.0-4.9 0.7 NSS 89.1
(0.7)
0.38 0.34-0.40 0.43 NSS 99.1
(0.14)
Reduction
Chrysotile
may exceed
99.4
98.3
may exceed
97.5
may exceed
87.2
may exceed
95.5
>99.1
may exceed
95.5
may exceed
97.2
Note: Detection Limit in Parenthesis
-------
TABLE 40. WATER QUALITY DATA - QUEEN LANE PLANT
00
M
Date
12/7-13/77
12/28/77-
1/3/78
2/22-28/78
11/13-19/78
1/8-15/79
Raw Water Filtered Water Percent Reduction
Turbidity Chrysotile Turbidity Chrysotile Turbidity Chrysotile
fibers fibers
Avg. Range Avg. Range
ntu ntu 106 f/L ntu ntu 106 f/L
4 2-10 2.9 0.11 0.06-0.23 0.2 97.2
8 4-24 5.5 0.20 0.17-0.22 0.1 NSS 97.5
(0.1)
3 3-4 3.6 0.06 0.05-0.08 (0.01) BDL 98
6 4-9 19.5 0.11 0.09-0.15 0.84 98.1
30 11-110 86 0.10 0.07-0.14 0.43 NSS 99.6
(0.14)
93.1
98.1
>99.7
95.6
may exceed
99.5
Note: Detection Limit in parenthesis
-------
TABLE 41. WATER QUALITY DATA - BELMONT PLANT
00
U!
Date
10/31/77-
11/6/77
12/14/77
Storm
12/21/77
Storm Rise
2/22/78-
2/28/78
11/13-19/78
1/8-15/79
11/13-14/79
settled water to
filter effluent
Raw
Turbidity
Avg. Range
ntu ntu
4.5 4-5
13 5-45
60 24-100
4 4
4.6 3.6-5.3
53 5-160
4.2 4.0-4.3
sand filter
Water
Chrysotile
fibers
106 f/L
7.6
7.6
14.6
2.5
8.6
15.2
10
Filtered
Turbidity
Avg. Range
ntu ntu
0.45
0.15
0.23
0.35
0.42
0.36
1.7
0.29
0.35-0.55
0.15
0.21-0.24
0.30-0.55
0.35-0.50
0.15-0.65
1.4-2.0
0.25-0.35
Water
Chrysotile
fibers
106 f/L
0.03 NSS
(0.01)
(0.01) BDL
0.02 NSS
(0.02)
0.03 NSS
(0.03)
0.28 NSS
(0.14)
(0.14) BDL
see data below
1.7
1.35
Percent
Turbidity
90
98.8
99.6
91.2
90.8
99.3
59.5
93.0
Reduction
Chrysotile
may exceed
99.6
>99.8
may exceed
99.8
may exceed
98.8
may exceed
96.7
>99.0
83.0
86.5
Detection Limit in Parenthesis
-------
SECTION 9
CHICAGO
The City of Chicago operates two very large water filtration plants, the
South District Filtration Plant rated at 480 mgd and the Jardine Water Filtra-
tion Plant rated at 960 mgd. Total peak filtration capacity of the two plants
is 2500 mgd. The peak filtration capacity of the Jardine Plant, formerly
called the Central District Filtration Plant, is 1,700 mgd, making this the
world's largest water filtration plant. The Jardine Plant was described in
American City (47) and in Water Works and Wastes Engineering (48). Both of
Chicago's plants are conventional rapid sand filtration plants filtering at 2
gpm/sf at rated capacity. Alum is used as the primary coagulant chemical.
Data on asbestos in raw and filtered water at Chicago have been obtained
in a long-range analytical program carried out by the Water Purification Labor-
atory of the City of Chicago. The report of McMillan, Stout and Willey was in-
cluded in the literature review (20). More recent data furnished by McMillan
are plotted in Figures 17 and 18, which show raw and filtered water asbestos
concentrations from January 1976 through December, 1978 for the Jardine and
South Plants, respectively.
The figures show that raw water generally contained 1 to 2 million fibers
per liter. Filtered water generally had 0.1 to 0.3 million fibers per liter.
Fiber reduction is similar to the percentage reported earlier (20), about 70
to 90 percent.
84
-------
(X
D Raw water
_ Filtered water
\
*^
10
o
z
o
QC 1,0
LU
O
2
O
U
CO
o
(-
to
LU
m
CO
0.1
.02
D
D n D
D
0°
D
D
D
J FMjAMJ JA,SONDJ FMAMJJ;ASONDJFMAMJJASOND
1976 1977 1978
Figure 17. Chicago asbestos monitoring data Jardine water filtration plant.
-------
00
ON
10
10
o
1.0
DC
Z
UJ
O
Z
o
o
o
H
UJ
03
C/5
0.1
DRaw water
Filtered water
DO
D 0
B
J FMAMJJASONDJFMAMJJASONDJ FMAMJ JASOND
1976 1977 1978
Figure 18. Chicago asbestos monitoring data South water filtration plant.
-------
SECTION 10
DISCUSSION
As the data for treatment plants in the literature review and results
sections indicated, some treatment plants have been very effective for fiber
removal, whereas others have not been effective. This section of the report
will present some possible reasons for the differences and suggestions for
plant operating techniques to enhance fiber removal.
MONITORING METHODS
Electron Microscope
The only method of analysis that provides positive identification of
asbestos fibers is the transmission electron microscope (TEM) method. Elec-
tron microscopy is both slow and expensive. Sample analysis generally takes
more than one working day, including all preparation steps, although analysis
of a single sample usually does not require more than eight hours of an ana-
lyst's time. Asbestos analysis by TEM generally costs $300 or'inore per
sample.
Because of the work load, most electron microscope laboratories (in 1979)
are not able to provide results in less than three or four weeks after receipt
of samples. Thus electron microscope results can not be used to monitor on-
going plant performance. They are very valuable, however, for reviewing past
plant performance to decide if treatment practice was effective for fiber re-
moval.
X-ray Diffraction
The measurements of amphibole mineral in Duluth's drinking water in 1973-
were made by X-ray diffraction (2). The method is quicker than TEM analysis
and costs about one-fifth to one-tenth of the electron microscope method.
Amphibole mass measurements were used during the Duluth pilot plant study and
have continued to be used by the EPA laboratory in Duluth.
One disadvantage of amphibole mass measurement by X-ray diffraction is
that the technique measures the mass of material and not the number of fibers.
The presence of some very large pieces of amphibole could greatly increase the
mass detected by the technique without increasing significantly the number of
fibers.
87
-------
X-ray diffraction has not been used to determine the mass of chrysotile
in any water supply with chrysotile contamination. Work has been done on the
measurement of chrysotile by X-ray diffraction, but many technical problems
must be resolved before this could provide a less-expensive method for detec-
tion of chrysotile in drinking water.
Particle Counters
A limited amount of work has been done with HIAC particle counters* in
efforts to learn if numbers of larger particles (cross-sectional area equal to
that of a sphere of 1.0 pm or 2.5 urn or larger) correlate with numbers of
asbestos fibers. Results are limited and inconclusive and sample contamina-
tion may exist. Results from Seattle, Everett, and Duluth suggest that an in-
line system with the filtered water piped directly to the counter would mini-
mize contamination and offer the best prospects for successful application of
this device.
Laser-Illuminated Optical Detector
Under a research grant from the U.S. Environmental Protection Agency, a
laser-illuminated optical particle detector was developed at the University of
Minnesota-Duluth. This device was described in an EPA report (49). A multiple
detector system uses differences in the scattering signatures of different
types of particulates to identify the particulates. This detector was used to
study red clay, taconite tailings, amphibole fibers, and chrysotile fibers.
Detector output was related to electron microscope counting data so that the
instrument could be used to estimate the concentration of fibers in water.
Fiber levels as low as 0.05 X 10^ f/L could be detected. An important advan-
tage of the detector was that results could be obtained in a matter of minutes.
Even when fiber counts are very low (under 0.01 X 10" f/L) results should be
available in less than one hour. This instrument provides a rapid estimate of
the number of fibers in water and is capable of identifying the types of parti-
cles present if the water does not contain particles that were absent when the
calibration was done.
Turbidity Measurement
Of the methods described in this section of the report, turbidity measure-
ment has perhaps the most disadvantages. Turbidimetry is not specific for
asbestos fibers and can not differentiate between light scattered by fibers or
clay or diatoms or any other particulates. Furthermore, Lawrence's data on
measurement of asbestos fibers show that fiber counts below 10-*-^ f/L are not
readily detected by a turbidimeter (27). Asbestos fibers in concentrations
found in most raw waters and nearly all filtered waters are too low to cause a
turbidimeter to register any light scatter. A turbidimeter alone can not be
used to detect and quantify asbestos fibers in drinking water.
Correlation of Turbidity and Fibers in Raw Water
Turbidity and asbestos fiber counts in unfiltered waters may correlate
weakly or very poorly. A weak relationship existed for Seattle raw water
chrysotile or amphibole and turbidity (Figures 19,20). Little or no relation-
ship was observed for amphibole fiber counts and raw water turbidity in the
*Pacific Scientific, Montclair, CA.
88
-------
25
20
o
O
s
00
vo
0
10
Point not used
LINEAR REGRESSION
C=1.36+6.1 (TURB)
R=0.66
Se=3.2x106
0.5
1.0 1.5 2.0
TURBIDITY, ntu
2.5
t
Points not used
3.0
From Kirmeyer (42)
Figure 19. Raw water chrysotile vs. turbidity Seattle pilot plant.
-------
5.0
^ 4.0
8
o
m
I
Q.
2.0
1.0
LINEAR REGRESSION
A=-0.5+2.5 (TURB)
R=0.67
Se=1.2x106
Points not used
0.5
1.0 1.5 2.0
TURBIDITY, ntu
2.5
3.0
From Kirmeyer (42)
Figure 20. Raw water amphibole vs. turbidity Seattle pilot plant.
-------
Duluth pilot plant study (Figure 21). These figures show why trying to asso-
ciate a fiber count with a given turbidity value often would not give a satis-
factory estimate of the asbestos concentration.
Turbidity Measurements for Evaluating Plant Operation
Turbidity measurements have been very useful for monitoring the filtration
process even though the fibers in the filtered water do not affect the turbidi-
meter reading. As a monitoring device the turbidimeter is used to assess the
overall efficacy of the granular media filration process,and this appears to be
related to asbestos fiber removal.
Seattle pilot plant filter run dataKirmeyer showed that turbidity meas-
urements could be used as an indicator of whether or not a granular media
filter was removing asbestos fibers (42).
Filter runs #21 and #24 were conducted using 7 and 9.2 mg/L of alum,
respectively, along with lime for pH control and a polymer as a filter aid.
Results are presented in Figures 22 and 23.
These results indicate that when finished water turbidity was _< 0.10 ntu,
then chrysotile counts were 0.34 X 106 f/L or less. At hour 7 of run #21 when
the finished water turbidity spiked at 0.34 ntu, the finished water chrysotile
rose from NSS levels up to 12.25 X 106 fibers/liter. High levels of chrysotile
also coincided with the abrupt rise in turbidity in run #24.
Filter run #174 was conducted using 10 mg/L of alum, lime for pH control
and a nonionic polymer as a filter aid. Initial filter loading rates were as
high as 10 gpm/ft2 and the declining rate filtration mode of operation was used.
To test the sensitivity of the turbidimeter in detecting changes in the opera-
tion of the filter, the alum feed pump was temporarily discontinued after the
first hour of operation and finished water turbidity was monitored continuously
during the process. Within about 15 minutes after the pump was disconnected,
the finished water turbidity began to rise rapidly. The detention time of the
water in the flocculation chamber and in the free space above the filter media
can be estimated at about 13.5 minutes. Although both chambers are complete
mix systems as opposed to plug flow reactors, a comparison of these two times,
15 vs. 13.5 minutes, indicates that the simple turbidimeter functions excep-
tionally well as a "troubleshooting" instrument that could have alerted an
operator that something was wrong somewhere in the system. Within 15 minutes
after the alum pump was placed back into operation, the finished water turbidity
returned to _< 0.10 ntu and stayed at that low level for the duration of the
filter run. ""Asbestos analyses conducted before, during and after pump shut-
down are presented in Figure 24. These data indicate that when the alum pump
was off and finished water turbidity had risen only to 0.20 ntu, a noticeable
increase in chrysotile counts occurred, as compared to the very low levels
previously measured.
Figure 25 provides an excellent example of how finished water turbidity
and asbestos counts track during filter breakthrough. This filter run indi-
cates that high chrysotile counts in the filtered water coincide with filter
breakthrough as indicated by a rising finished water turbidity.
91
-------
LLJ -f
£s
II
4.0r-
it 30
o9
n
2.0
1.0
l
1.0 2.0 3.0
amphibole count, 106f/L
J
4.0
From Black & Veatch (12)
Figure 21. Relationship between raw water turbidity at Duluth
Lakewood intake and ORF amphibole fiber counts.
-------
1.2
3
c
. 1.0
j
Q
CO
3 0.8
CC
LU
h-
£ 0.6
Q
**o LU
w r
CO
Z . -
_ 0.4
LL
0.2
0
T
|| RAW WATER TURBIDITY = 0.70 ntu
' \ RAW WATER CHRYSOTILE COUNT = 25. 8x1 08 f/L
i
I
i \
' I
I \
\
I '
i
| \
1 1
1 i
I .
1 i
i \
1
'
f 1
1 1
| « |
1 / \ |
1 / \ \
' / V / TURBIDITY
i / \ 1 /
'/ \l /
^r* *"T """"^H " i "*" 3 1 I
2 4 6 8 10 12 14 16 18
12 _i
^
O
fc
10 t
z
D
O
O
LU
8 d
h-
O
(0
DC
6 o
CC
LU
1-
4 ^
Q
LU
X
0)
2 ?
LL
TIME, hours
Figure 22. Finished water turbidity and chrysotile vs. time-run
Seattle pilot plant.
From Kirmeyer (42)
-------
1.2
3
+>
C
2 0.8
m
cc
D
oc
UJ
I 0.4
CO
RAW WATER TURBIDITY - 0.60 ntu
RAW WATER CHRYSOTILE COUNT = 9.4x106 fibers/liter
4
Chrysotile
Turbidity
«*.
m
8 10 12
TIME, hours
14
16
18
12
8
-------
0.5r 1-0
0.4
*-
I \
Q.
O)
6 ^J
<
cc.
5 O-
Q
4 §'
3 pj
i
2
1
10
(O
O
_j
Q
<
ui
X
6 8 10 12 14 16 18 20 22 24
TIME, hours
From Kirmeyer (42)
Figure 24. Operating data for run #174MM Seattle pilot plant.
-------
1.2
1.0
<£>
o
0)
cc
UJ
ffi
0.8
0.6
DC
I
O 0.4
0.2
RAW WATER TURBIDITY - 3.3 ntu
1.2
1.0
3
c
*
>-
0.8 |
cc
D
0.6
cc
UJ
h-
Q
111
CC
LLJ
0.4
0.2
0 2 4 6 81012141618202224
TIME, hours
Figure 25. Finished water chrysotile counts and turbidity
vs. time-run #120 Seattle pilot plant.
From Kirmeyer (42)
96
-------
j^urbidity-fiber count relationship in filtered waterThe nature of the
turbidity-fiber relationship for filtered water could perhaps be compared to
the coliform-pathogen relationship. The presence of coliforms is a signal that
pathogens might be present. It is, however, not certain proof of pathogen
presence and no direct proportional ratio for the two kinds of microorganisms
exists. Similarly, when a raw water contains asbestos fibers, a rise in fil-
tered water turbidity very probably is a sign of an increase in filtered water
fiber count, but the fiber count can not be estimated on the basis of the
abnormally high turbidity reading. Data from the Seattle study are presented
in Table 42 to illustrate this. One sample with 0.24 ntu turbidity had
11 X 106 f/L, whereas another with 0.28 ntu had a not statistically significant
count of about 0.14 X 106 f/L.
When a granular media filter has been operating for a period of time and
has been producing a stable quality of effluent as measured by turbidity, a
rise in filtered turbidity should be interpreted as a sign that some operator
action is required. This may mean an adjustment of chemical doses or a filter
backwash. When a granular media filter is removing a contaminant in particu-
late form, such as asbestos fibers, a rise in filtered water turbidity is a
strong indication that the particles that had been captured within the filter
are now being released.
Table 43 contains data on fiber counts during Seattle filter Runs 53 and
62. In Run 53, when the turbidity rose after nine hours of filter operation,
the chrysotile count in the filtered water was nearly four times that in the
raw water. In Run 62, when turbidity rose to 0.37 ntu at 16 hours, chrysotile
counts in raw and filtered water were quite similar. Amphibole data for Run
62 giveadditional evidence that granular media filters can store and then
slough particulate contaminants. Amphibole fibers were not detected in raw
water or in filtered water during normal operation in this run, but when tur-
bidity rose at hour 16, amphibole fibers that apparently had been removed and
concentrated in the filter were released in a concentration high enough to be
detected in the filtered water.
The sloughing and discharge of asbestos fibers from filters that prev-
iously had removed them from the raw water is a potential problem that filter
plant operators must consider. At the present time, the best way to prevent
the discharge of a concentrated amount of fibers into filtered water seems to
be to avoid any significant rises in turbidity after a granular media filter
has ripened and effluent turbidity is stable.
The extent of a turbidity rise that could cause a problem is not well
defined, but in Run 53 an increase from 0.13 ntu to 0.24 ntu resulted in a
forty-fold increase in the chrysotile count. For a filter producing 0.10 ntu
filtered water, a rise of 0.1 ntu is probably sufficient cause to backwash the
filter. Controlling the operation of a filter in this manner is being accom-
plished at some filtration plants, but for many others it would represent a
considerable modification of operating procedures.
97
-------
TABLE 42. SEATTLE PILOT PLANT CHRYSOTILE FIBER COUNTS IN FILTERED WATER
WHEN TURBIDITY ROSE ABOVE 0.10 ntu
Run # Hours Into
Run
21 7
24 7
53 9
62 16
120 1
120 19
120 20
120 21
120 22
120 23
Filtered
Turbidity
ntu
0.34
0.36
0.24
0.37
0.14
0.14
0.28
0.48
0.57
1.2
Fibers
106 f/L
12.25
6.2
11.2
2.28
0.14
0.19
0.14 (NSS)
0.28
0.57
1.25
TABLE 43. RELEASE OF ASBESTOS FIBERS BY FILTERS AT SEATTLE PILOT PLANT
Run
Sample Hour Into Turbidity
Location Run ntu
Chrysotile
fibers/liter
Amphibole
fibers/liter
53
53
53
62
62
62
62
RAW
FINISHED
FINISHED
RAW
FINISHED
FINISHED
FINISHED
9
3
9
16
3
9
16
0.50
0.13
0.24
0.35
0.1
0.105
0.37
3.0 X 106
0.27 X 106
11.2 X 106
2.52 X 106
0.34 X 106
0.24 X 106
2.28 X 106
0.70 X 106 (NSS)
<0.02 X 106 (BDL)
1.2 X 106 (NSS)
<0.1 X 106 (BDL)
<0.02 X 106 (BDL)
<0.02 X 106 (BDL)
0.1 X 106
From Kirtneyer (42)
98
-------
GRANULAR MEDIA FILTRATION FOR ASBESTOS FIBER REMOVAL
The ability of granular media filters to remove asbestos fibers has been
demonstrated repeatedly in filtration experiments at Duluth and Seattle.
Operating data from the water treatment plants at Duluth, Two Harbors, and
Silver Bay confirm pilot plant studies that showed that the fibers can be
removed on a routine, day-to-day basis. Because a properly operated filter
can reduce the concentration of asbestos fibers in water by as much as 99 to
99.9 percent, or even more in some instances, the problem of efficiently
removing fibers from water becomes a problem of achieving proper filter
operation. In this section of the report, operating techniques are dis-
cussed. Certain aspects of filter plant design were shown to be important
in the pilot plant studies, and these are also presented.
Filtered Water Turbidity Value for Effective Fiber Removal
Turbidimeters are used to monitor filter performance in many water treat-
ment plants. Because turbidity changes are associated with changes in the
asbestos fiber concentration of filtered water under certain circumstances,
turbidity and asbestos fiber count data have been analyzed in the search for a
relationship between filtered water turbidity and fiber concentration. Such a
relationship was found for the Seattle pilot plant and for the water filtra-
tion plants on the north shore of Lake Superior. Because of the differences
in fiber counting, the 1974 Duluth pilot plant data are analyzed and dis-
cussed separately from data collected since January, 1977.
Filtered Water Turbidity vs. Fiber Count
Duluth amphibole dataThe Duluth pilot plant study (12) showed that
amphibole fiber counts were likely to be at or near the detection limit when
filtered water turbidity was below 0.2 ntu. In Figure 26, of the 41 BDL counts
reported, 31 occurred when filtered water turbidity was 0.10 ntu or lower.
Logsdon and Symons (33) recommended that quality control for filter operation
at Duluth should be based on producing finished water with a turbidity of not
more than 0.10 ntu with the run being terminated when the turbidity reached
0.2 ntu.
Seattle amphibole dataOne of the goals of the Seattle pilot plant
study was to confirm the Duluth findings. To do this the Seattle Water Depart-
ment focused upon turbidity and fiber removal. A summary of the results of an
extensive program of testing, sampling, and analysis is presented in Appendix
B. This appendix gives turbidity and fiber count data for the pilot plant
runs that were sampled for electron microscope analysis.
Review of the Seattle fiber removal data indicates that amphibole fibers
can consistently be removed down to the detection limit of 0.01 X 10" fibers/
liter. Removal efficiencies were between 98.3 and 100%. When amphiboles were
detected in the raw water and when turbidity was < 0.10 ntu, 52 out of 57
results (91%) had levels of amphibole which were at or less than the detection
limit. In all 5 of the cases where amphibole fibers were > 0.01 X 10" fibers/
liter, 4 or fewer fibers were counted by the analyst, which means that the
results were not statistically significant. Three out of the five cases
99
-------
1.0r
0.3
CO
o
V)
oc 0.1
LU
03
LLI
_1
O
m
0.03
0.01
D
Tf
D
m n ..p a
1 10 5 5 1 1 1 1 1
a
NOTE: The number below a data point represents
the number of times a given turbidity
and BDL fiber count occurred.
I
0.1
0.2
0.3
0.4
0.5
FILTERED WATER TURBIDITY, ntu
Figure 26. Effluent turbidity vs. amphibole fiber count, Duluth granular
media pilot plants.
100
-------
occurred when alum without lime addition was being employed as a treatment
method and the other two occurred at an alum dosage of about 7 mg/L rather
than the optimal 10 mg/L. These are possible explanations for the outlying
data points but by no means are they conclusive evidence. As indicated in
Figure 27, when the finished water turbidities exceed 0.10 ntu, amphibole
counts are much higher and the data show a noticeable scatter. These data
points represent times when either poor destabilization was occurring or when
turbidity breakthrough had already taken place. In conclusion, amphibole
fibers can consistently be removed down to the detection limit if certain oper-
ating conditions are met. This confirms the earlier amphibole research and
operating data at Duluth, Minnesota (12).
Seattle chrysotile data Review of the chrysotile data listed in Appen-
dix B indicates that although excellent removals can be achieved, it is more
difficult to remove than amphibole and results are more variable. Figure 28
shows a very tight pattern of data when finished water turbidity is _< 0.10 ntu
and considerable scatter above that point. As with the amphibole results, the
scattered chrysotile data above 0.10 ntu are normally associated with either
poor destabilization, turbidity spikes or breakthrough. Both Figures 27 and
28 are very similar to results obtained in the Duluth pilot plant study.
Turbidity and Probability of High Fiber Counts
Lake Superior results In the operation of the Lakewood Filtration Plant
at Duluth, filtered water turbidity for plant monitoring samples was routinely
0.04 to 0.06 ntu. The lower turbidity values approach the detection limit for
the nephelometer employed, and asbestos results are frequently at or near the
detection limit too. A direct comparison of filtered water turbidity vs.
fiber count would not be very useful, so another approach was taken (13).
Filtered water fiber count data from Duluth were grouped according to turbidity
and operating conditions. Differences in the fiber counts for the data groups
are apparent in Figure 29, which shows the probability of occurrence of fiber
counts for the various operating conditions. Lowest fiber counts were likely
to occur during routine operating conditions when turbidity was below 0.10 ntu.
Highest counts were likely to occur when filter rate changes took place and
turbidity was 0.10 ntu or greater.
In this figure and in Figures 30 and 31, when a BDL fiber count was re-
ported, the data point was plotted at the detection limit. This resulted in
clusters of data points at certain fiber count values and is a conservative
approach to data analysis, because actual fiber counts could have been below
the BDL values.
In this report an effort is made to extend this comparison to the Silver
Bay plant. Data from this plant are included in the probability plots of
Figure 30. In this figure, data are divided according to filtered water
turbidity (equal to or below 0.10 ntu vs. greater than 0.10 ntu), and
amphibole fiber counts were tabulated in order. Figure 30 shows the proba-
bility of occurrence of fiber count according to the circumstances of
sampling as well as filtered turbidity. The samples likely to have the high-
est fiber counts were those from the new 10 gpm pilot plant with turbidity
above 0.10 ntu. Samples from the treatment plants with turbidity above 0.10
ntu can be seen to have a likelihood of higher amphibole counts than samples
101
-------
1.0 -
0.3
ID
o
«- 0.1
CC
LJJ
m
LL
LJJ
g O-O3
x
Q.
0.01
Region of most successful removal
G C
t
a g a
23
DO
D D
NOTE: The number below a data point
represents the number of times
a given turbidity and fiber
count occurred.
0.1 0.2 0.3 0.4
FINISHED WATER TURBIDITY, ntu
0.5
From Kirmeyer (42)
Figure 27. Amphsbole count vs. finished water turbidity Seattle
pilot plant.
102
-------
10.0 -
3.0
1.0
1C
o
to
DC
UJ
CD
O
c/j
i
o
0.3
0.1
0.03
0.01
Region of most successful removal
n
ODD
a
n a
n n n n I
41896,
NOTE: The number below a data point represents
the number of times a given turbidity
and fiber count occurred.
'0,1 0.2 0.3 0.4 0.5
FINISHED WATER TURBIDITY, ntu From Kirmeyer (42)
Figure 28. Chrysotile count vs. finished water turbidity Seattle
103
-------
10
-------
100
30
10
<0
O
CO
oc 1.0
ui
m
LL
UJ
m 0.3
o.
2
<
0.10
0.03
0.01-
TURBIDITY
>0.10 ntu
^ 0.10
A ^
> 0.10 ntu
0.10 ntu
PLANT
New pilot plant
at Duluth
o
Treatment plants
at Duluth, *
Silver Bay
o
o
o
00<>
o
^kAV>v
9 samples
12 samples
* 7 samples
r6 samples
4A
^9 samples
t A
26 samples
A A^A>
1
'f -
34 samples
i iii i
i
10 30 50 70 90
99
SAMPLES WITH FIBER COUNT EQUAL TO OR BELOW SPECIFIED
VALUE, percent
Figure 30. Frequency distributions of amphibole fiber counts
for filtered waters from plants on Lake Superior.
Data base January 1977 thru June 1979
105
-------
10
\
VH-
cc
I
O
1.0
0.1
0.01
Filtered Water Turbidity
A ^0.10 ntu
& above 0.10 ntu
A A
AAA
38 fiber count observations
S 0.01x106f/L
I i i
1 10 30 50 70 90 99
SAMPLES WITH FIBER COUNT EQUAL TO OR BELOW
SPECIFIED VALUE, percent
Figure 31. Frequency distribution of chrysotile fiber counts
for filtered water sample groups Seattle pilot plant.
106
-------
taken at the treatment plants when turbidity was 0.10 or less. The high
turbidity monitoring data consist of two samples taken at Duluth when an
upset of coagulation chemistry caused very high fiber counts. These samples
were not taken during a special rate change test so they were included as
high turbidity monitoring samples.
Figure 30 clearly shows that lower fiber counts are more likely to
occur when the turbidity of filtered water is 0.10 ntu or lower. In this
turbidity range, amphibole fiber counts were 0.03 X 10^ f/L or less half
of the time for the filtration plants. For this reason, the recommendation
is made that water filtration plants treating Lake Superior water should
attempt to produce filtered water meeting a goal of 0.10 ntu.
Seattle resultsKirmeyer (42) showed a similar trend with Seattle chryso-
tile data. Figure 31 can be used to estimate what levels of chrysotile were
present at any given time when finished water turbidity was _< 0.10 ntu or above
0.10 ntu. For example, 50 percent of the time when turbidity was £ 0.10 NTU,
finished water chrysotile counts were _< 0.03 X 106 fibers/liter. When turbid-
ity was > 0.10 ntu, 50 percent of the time, chrysotile counts were _< 0.3 X
10° fibers/liter. Thus, over 10 times more asbestos was present in the
finished water when turbidity was > 0.10 ntu as compared to when turbidity
was _< 0.10 ntu. In conclusion, chrysotile fibers can consistently be
removed to levels which are NSS if certain operating conditions are met.
As a further test of the relationship of fiber removal and filtered water
turbidity, a statistical analysis was performed using Seattle data in Appendix
B. Fiber removal percentages were calculated, and data were placed into four
fiber removal categories: > 99.0 percent, > 95.0 to 99.0 percent, > 90.0 to
95.0 percent, and 90.0 percent or less. Then the number of each of these
samples with filtered water turbidity above 0.10 ntu and 0.10 ntu or less was
placed in a table (Table 44). This was done separately for amphibole and chry-
sotile results. The hypothesis in each case was that the frequency of observa-
tion of the various fiber removal percentages was independent of filtered water
turbidity. The results were statistically different from the expected (null
hypothesis) at the 0.01 level for both amphibole and chrysotile. Thus the
Seattle results also showed that the degree of fiber removal is related to
filtered water turbidity.
Filtered Turbidity and Fiber Reduction at Seattle
In another test of the relationship of fiber removal and turbidity, fiber
removal data for chrysotile were placed into the four percentage categories
used in the chi square analysis. Then data for each category were arranged
in order according to filtered water turbidity. The results were plotted on
logarithmic probability paper, showing the percentage of samples of a given
fiber removal having turbidity equal to or less than a certain value. Results
are shown in Figure 32. This figure clearly shows that higher filtered water
turbidities are associated with lower percentages of chrysotile removal.
Because the relationship shown in Figure 32 is very important, a careful
explanation of what it does show and what it does not show is appropriate. The
figure does not show that the percentage of chrysotile removal is always higher
when turbidity is lower, because this is not what happened in the pilot plant
107
-------
TABLE 44. CHI SQUARE ANALYSIS OF FIBER REMOVAL AND TURBIDITY - SEATTLE
Chrysotile Data
Turbidity of Fiber Removal Percentage Categories
Filtered > 99.0% > 95.0 to 99.0% > 90.0 to 95.0% _< 90.0%
Water Number of samples with fiber removal percentage in category
_< 0.10 ntu 48 43 73
> 0.10 ntu 2 7 6 9
X2 = 38.5
Amphibole Data
Turbidity of Fiber Removal Percentage Catagories
Filtered > 99.0% > 95.0 to 99.0% > 90.0 to 95.0% _< 90.0%
water Number of samples with fiber removal percentage in category
_< 0.10 ntu 6 9 0 0
> 0.10 ntu 00 12
X2 =18.0
study. For example, one filtered water sample with a removal percentage of 90
percent or less had a turbidity of 0.06 ntu. Some fluctuation occurs from time
to time, partly because of the variation in analytical methods and partly be-
cause of the variation in treatment efficacy. Figure 32 shows the trends that
occur as many samples are taken over a period of time. It shows the distribu-
tion of filtered water turbidities for a given level of fiber removal.
An example of how the figure could be used follows. If, on the basis of
raw water data, a water department chose to operate a filtration plant to
attain at least 99 percent reduction in the fiber count, producing water with
a filtered turbidity of 0.10 or less would be necessary 95 percent of the time.
Filtered water turbidity, over a long period of time, could be allowed to
exceed 0.10 ntu only 5 percent of the time in order to consistently attain 99
percent fiber reduction. On the other hand, if only a 90 percent reduction of
chrysotile was needed, turbidity would have to remain below 0.35 ntu 95 percent
of the time. If chrysotile removal was of no consequence, and less than 90
percent was acceptable, meeting the 1 ntu limit of the Drinking Water Regula-
tions could be an operating goal.
108
-------
3
+
C
1.0 -
g
QQ 0.5
cc
D
r-
cc
LLI
<
g 0.2
CC
HI
0.10
0.05
Percentage chrysotile fiber
count was reduced
0«90.0%
| >90.0 to 95.0%
A > 95.0 to 99.0%
>99.0%
1 10 30 50 70 90 99
SAMPLES WITH FILTERED TURBIDITY EQUAL TO OR LESS THAN GIVEN VALUE, percent
Figure 32. Relationship of turbidity and chrysotiSe removal by
granular media filtration Seattle pilot plant.
-------
Aspects of Plant Design and Operation
Adequate Pretreatment Unit Processes
Water must be properly conditioned before filtration if the filter is to
achieve its maximum treatment efficiency. The need to prepare water for fil-
tration is generally acknowledged by sanitary engineers, and much research has
been conducted on rapid mix, flocculation, and sedimentation. Some aspects of
pretreatment were investigated in filtration research on asbestos fiber re-
moval.
Rapid mixingRapid mix variations were tried at Duluth, and the three-
stage rapid mix at the filtration plant was designed on the basis of pilot
plant studies. Rapid mixing was studied in greater depth at Seattle. Ade-
quate head is available at Seattle, so in-line mixers were used in the
pilot study. A comparison of back mixers and in-line or static mixers was
made. The usual cause for run termination with static mixers was excessive
head loss, whereas use of back mixers resulted in run termination because
of turbidity breakthrough. Because control of turbidity breakthrough is
very important, and head is available to permit use of static mixers,
in-line or static mixers would probably be used if a full scale plant is
built at Seattle.
FlocculationThe effect of flocculation was studied in Seattle by con-
ducting a series of filter runs with and without the flocculator. The results
in Table 26 showed that flocculation greatly increased the length of filter
runs, in terms of operating hours when the water quality goal was met. Pre-
conditioning by both rapid mixing and flocculation resulted in production
of lower filtered water turbidity. Some runs in the rapid mix-filtration
mode had no filtered water with turbidity equal to or below 0.10 ntu, and
so had no hours of production of acceptable water quality.
Similar results with Cascade Mountain water were observed in the Everett
pilot plant study (31). Watkins et. al. reported that when raw water turbid-
ity was high (10-20 ntu in this case) filtration preceeded by rapid mixing
and flocculation (direct filtration) was more effective than filtration
preceeded by rapid mix only (in-line filtration).
SedimentationSedimentation for asbestos fiber removal has been given
little consideration because the pilot plant studies conducted so far have
been conducted with clear water sources that did not require sedimentation
for effective filtration. Sedimentation was studied to a limited degree
at Everett (31), and the Duluth filtration plant can use this process, and
in fact does so a majority of the time.
Because the purpose of sedimentation is considered to be the same as that
of pretreatment processes for direct filtration - preparation of raw water for
filtration - emphasis in asbestos sampling has been on raw and filtered water
quality. Only limited sampling to evaluate asbestos removal in the sedimenta-
110
-------
tion process has been done. Canadian pilot plant studies (29) showed as much
as 60 to 80 percent removal of spiked chrysotile in the sedimentation process.
Two sets of samples collected at Philadelphia show 60 and 90 percent reduction'
of chrysotile. Philadelphia data include the effect of storage in presedimen-
tation basins at the Torresdale and Belmont plants.
Even though primary concern is on filtered water fiber count, the amount
of asbestos fibers removed in sedimentation vs. granular media filtration may
be of interest. For example, decisions on treatment, disposal or recycle of
backwash water would involve consideration of asbestos fiber levels in the
wash water. Decisions on water plant sludge disposal may be affected by
concern for the fiber content of the sludge and the need to avoid recontam-
inating the environment with the fibers after they have been removed from
drinking water. The fate of coagulated fibers in filter washwater or sedimen-
tation basin sludge is not clearly defined at this time.
Process Control
In order for the necessary treatment processes, described in the previous
section, to remove asbestos fibers at maximum efficacy^ they must be operated
properly. Careful control of water chemistry is essential for attaining effec-
tive coagulation and filtration. Two key elements of effective coagulation
are control of pH and use of the correct doses of coagulants and filter aids.
The optimum combination of treatment chemicals will vary from plant to plant
as well as from time to time at one plant. Techniques for determining
appropriate chemical doses have been discussed in the literature for many
years. This report will consider how process control affects fiber removal.
pH controlThe necessity for careful control of pH was shown by Kirmeyer
(42). Figure 8 shows the effect of pH on filtered water turbidity and alumi-
num content when the primary coagulant was alum. Kirmeyer concluded that pH
should be between 6.1 and 6.7 pH units to treat Tolt Reservoir water and
meet the goals of 0.05 mg/L of aluminum and 0.10 ntu turbidity. He stated,
"... pH is a very critical factor in the destabilization of Tolt water." (42)
Control of pH is also very important for plants treating Lake Superior
water. At Duluth the optimum pH for treatment is influenced by the raw water
temperature, ranging from 7.2 or 7.3 at 1°C to 7.05 at 7°C and 6.8 at 13°C.
When sedimentation is employed, a deviation of + 0.1 pH unit from the optimum
value is acceptable. For filtration without sedimentation, plant operators
try to maintain pH within + 0.05 unit of the optimum value.
Use of adequate treatment chemical doseIn addition to the need for care-
ful control of pH during coagulation, another very important aspect of treat- v
ment process control is maintaining the proper doses of treatment chemicals.
Polymers generally have a range of concentration that gives optimum efficacy,
with reduced effectiveness resulting from either an overdose or an underdose
of coagulant aid or filter aid. Inorganic coagulants must be used in doses
adequate to assure that particulates are destabilized before filtration.
When very clear water (close to 1 ntu) is filtered, meeting the 1 ntu
Maximum Contaminant Limit for turbidity can sometimes be accomplished without
111
-------
the addition of coagulant chemical. Often 1 ntu filtered water can be produced
with a very small coagulant dose, such as 2 to 5 mg/L of alum. Producing very
low turbidity filtered water, 0.10 ntu or below, requires higher doses of
treatment chemicals and more careful filter operation.
Data available now suggest that the turbidity-asbestos relationship holds
only for waters treated with enough coagulant either to agglomerate particles
so they will settle or to destabilize all or nearly all of the particles in
the raw water so they will be removed in the filter. San Andreas plant data
(Table 35) show that when very low coagulant doses or no coagulants were used,
fiber counts were in the range of 1 to 2 X 106 f/L even though filtered tur-
bidities were 0.1 to 0.3 ntu. Kirmeyer (42) also reported that use of a low
alum dose (6 mg/L) resulted in filtered water turbidity of 0.20 ntu but a
chrysotile count of 1.6 X 10^ f/L. Apparently filtration plant operators try-
ing to maximize asbestos fiber removal should not try to minimize application
of coagulant chemical.
A review of Chicago data suggests that the use of low doses of coagulant
chemical can result in the removal of a large portion of the asbestos fibers
(70 to 90 percent usually), but the relationship between turbidity and fiber
count is weak or non-existent when less than optimum coagulant doses are used.
Since 1976, typical coagulant use at the Jardine Water Filtration Plant has
been in the range of 2 to 8 mg/L of alum plus 0.2 to 0.4 mg/L of cationic
polymer as a coagulant aid. At the South Water Filtration Plant typical alum
doses range from 2 to 8 mg/L, with 0.1 to 0.3 mg/L of cationic polymer gener-
ally used. These doses are not now nor ever have been designed specifically
for asbestos fiber removal and are lower than doses of chemical used success-
fully for asbestos removal at Lake Superior (10 to 15 mg/L of alum and 0.05 to
0.1 mg/L of nonionic polymer). Chemical doses employed at Chicago result in
production of filtered water that has a turbidity of 0.15 to 0.25 ntu most of
the time, although occasional samples may be as low as 0.10 ntu or as high as
0.35 ntu. The turbidity limit set by Chicago was intended to keep the dis-
tribution system clean, and this has been in effect for two decades.
That the chemical doses used are not intended to destabilize all of the
particles in the raw Lake Michigan water is possibly shown by the variation of
filtered water turbidity. It is also evident in the lack of chrysotile counts
below detectible limits (BDL) at Chicago. BDL counts were frequently observed
at Seattle and Duluth, where the higher coagulant chemical doses used resulted
in lower filtered water turbidity. Lake Michigan water at Chicago is at times
as clear as Lake Superior (1 ntu), but the Chicago filtration plants use lower
doses of coagulant chemicals than filtration plants at Duluth, Two Harbors and
Silver Bay.
The differences in chemical use at the two locations are related to vari-
ations in raw water quality, water quality goals, and economics. Because of
the low fiber count in filtered water at Chicago (0.1 to 0.3 X 106 f/L), and
because EPA has not established any limit for asbestos fibers in water, man-
agement in Chicago does not now (Fall, 1979) see any need or reason to concern
itself about asbestos removal. In contrast, treatment plants on Lake Superior
must treat water with amphibole fibers often in the 10 to 100 X 10^ f/L con-
centration range so deliberate efforts are made to maximize fiber removal.
112
-------
The turbidity goal at Chicago, 0.2 ntu, is well below the 1 ntu Maximum Con-
taminant Level for turbidity, but above the 0.10 ntu goal recently estab-
lished for Duluth, Silver Bay and Two Harbors. Filtered water with 0.10
ntu turbidity could be produced at Chicago. However, because of the
large and presently difficult to justify extra cost for chemicals this is
not done.
The Chicago monitoring data show that coagulant doses that produce water
that always meets the 1 ntu turbidity MCL may not achieve maximum asbestos
fiber removal.
Filter Design
A number of factors related to filter design have been observed or eval-
uated with respect to fiber removal. These include type of filter media, rate
of filtration, type of rate control, and depth of water over the media. These
are discussed in this section of the report.
MediaThe various kinds of granular media generally used for water fil-
tration have been shown to remove asbestos fibers. Conventional rapid sand
filters operated at 2 gpm/sf (like those at Philadelphia) are seldom chosen by
design engineers, because of the advantages and benefits of filtration with
dual or mixed (tri) media. Therefore, efforts to evaluate filter media have
been concentrated on dual and mixed media.
Dual and mixed media were compared in the 1974 Duluth pilot plant study.
When Black & Veatch designed the filtration plant at Duluth, mixed media was
selected. Dual media was placed in one of the four filters at the Lakewood
plant, however, so a full-scale comparison could be made. Results suggest
that in instances when filtered water quality deteriorated as a result of a
filtration rate change, the quality of the mixed media filter effluent was
generally better than the quality of the dual media filter effluent.
Kirmeyer compared dual and mixed media for both asbestos fiber removal
and water production efficiency in Seattle. Comparisons for fiber removal
involved both coarse and fine dual and mixed media. Details were given by
Kirmeyer (42). In the fiber removal evaluation, two sets of three simultan-
eous filter runs were carried out, and 45 water samples were analyzed for
asbestos. Kirmeyer concluded that under normal operating conditions, all of
the different media tested were equally effective for asbestos fiber removal.
Concerning water production, Kirmeyer found little difference in net
water produced per day for dual media and mixed media in a range of filtration
rates from 5.5 gpm/sf to 7.5 gpm/sf. Data on water production efficiency are
likely to be specific for the raw water source, so pilot plant tests with
local water would be needed before a decision could be made on media types and
water production efficiency at other locations.
Rate and rate controlDuring the 1974 pilot plant study in Duluth, the
filters were operated in a flow range from 2 to 8 gpm/sf, but only a limited
number of runs with rates above 4 gpm/sf were sampled for asbestos fibers. On
the basis of the limited data, Black & Veatch concluded that filtration at 5
gpm/sf would be effective for amphibole fiber removal (12). In the Seattle
113
-------
pilot plant study, filtration rates as high as 10 gpm/sf were found to be
effective for asbestos fiber removal.
At Seattle, filters were operated in both declining rate mode and con-
stant rate mode. Both modes were effective for asbestos fiber removal.
Selection of type of rate control for a treatment plant would probably be in-
fluenced by factors such as experience of the design engineer with variable
and constant rate control and willingness of state regulatory engineers to
approve variable rate filtration. One possible advantage of variable rate
filtration is that abrupt rate changes are avoided. As noted before, Duluth
studies of rate changes showed that fiber counts could increase during abrupt
rate increases so these disturbances should be prevented to the extent pos-
sible.
Air binding of filtersIn pilot plant studies at both Duluth and Seattle,
air binding of the filters occurred, as indicated by air bubbles rising out of
the filter media when the filters were shut off at run termination. In both
locations the raw water was cold and nearly saturated or supersaturated with
dissolved gases. These can be released during filtration, especially when
water depth over the media is only 2 or 3 feet and the filter is operated to 7
or 8 feet of head loss.
Because of the air binding problem in 1974, the filtration plant at
Duluth was designed to prevent this. The filter overflow is 9 feet above the
top of the media so that 7 feet of water normally will be over the top of the
media during filtration. This design feature has prevented repetition of
the air binding problems that occurred in the pilot plant study.
Influence of Source Water Turbidity on Treatment Results
The source waters at Seattle and along the north shore of Lake Superior,
where extensive filtration studies have been carried out, are very clear and
generally unpolluted waters (except for the asbestos fiber content). Filtra-
tion of very clear water requires diligent operator attention if the treatment
process is to attain maximum efficiency. Particles must be completely destabi-
lized and low filtered water turbidity must be achieved to remove asbestos
fibers.
When turbid raw water is treated, the presence of a large number of parti-
culates in the water seems to aid the agglomeration of particles, and some
water plant operators have observed that turbid waters may be easier to treat
than very clear waters. Some of the data presented in this report were ob-
tained from filtration plants that treat water sources that are not as clear
as Lake Superior and the Tolt Reservoir. These include a number of surface
waters in the San Francisco Bay area and the water sources for Philadelphia.
Turbidity for some of these was as low as 3 to 5 ntu during some sample
times. Turbidities in excess of 100 ntu were also observed. In contrast
with the clear (equal to or less than 1 ntu) lake waters, these sources
are considered turbid.
The relationship between filtered water turbidity and fiber count may be
different in waters that are more turbid than Lake Superior and the Tolt
114
-------
Reservoir. For example, in Contra Costa County, the Antioch and Pittsburg
plants had effluent turbidities between 0.1 and 0.4 ntu (Tables 32 and 34)
while effluent turbidity at the Bollman Plant was 0.06 ntu or lower (Table 33).
Nevertheless fiber counts in filtered water were similar for all three plants,
below 0.3 X 10° f/L in 12 of 14 instances. Two of these samples were in the
0.7 - 0.8 X 10^ f/L range, and six samples had BDL or NSS values. Raw water
turbidities generally were between 20 and 40 ntu.
At the San Geronomo plant (Table 36) two samples had counts above
1 X 106 f/L (2 X 106 and 12 X 106). These samples had filtered water turbidi-
ties of 1.0 - 0.6 ntu and 0.6 - 0.3 ntu. Raw water turbidity at the San
Geronomo plant was 5 ntu for one of the above samples and 46 ntu for the other.
All three plants at Philadelphia effectively remove asbestos fibers,
generally to NSS or BDL levels, as shown in Tables 39 - 41. This occurred
even though average filtered water turbidity varied from 0.06 to 0.20 ntu, at
Queen Lane, from 0.13 to 0.29 ntu at Torresdale, and from 0.15 to 0.45 ntu at
Belmont. Raw water turbidities were as low as 3-4 ntu and as high as 160 ntu.
The results from Philadelphia, Contra Costa County, and Marin County
suggest that attaining a filtered water turbidity of 0.10 ntu or less is not
as important when the raw water is not exceptionally clear, meeting the 1 ntu
MCL, before filtration. Perhaps when raw water contains enough clays, algae,
bacteria, and other particulates to raise the turbidity significantly above 1
ntu, the asbestos fibers become associated with the other particulates and are
removed along with them. This might occur especially with chrysotile, which
has a positive zeta potential at usual raw water pH values.
Data suggest that asbestos fiber removal can be accomplished when fil-
tered turbidity exceeds 0.10 if the raw water turbidity is greater than "1 ntu.
Sufficient data to define "turbid" waters for which this is true have not been
obtained, however, nor is the upper limit for filtered water turbidity that is
associated with good fiber removal apparent. Limited data suggest that 0.4
ntu might be adequate, but 0.6 ntu might not.
Too little data and too much uncertainty exist now to suggest an operat-
ing goal other than 0.10 ntu, or the AWWA Quality Goal of 0.1 (no trailing
zero) ntu. The AWWA goal technically can be met with a turbidity of 0.14 ntu
because of rounding. Treating to attain the AWWA Quality Goal or 0.10 ntu
would require careful operation and excellent equipment performance at plants
treating turbid water. A distinct advantage of such an exacting quality goal
is that to achieve it, fluctuations in filtered water quality would not be
permitted. A rise in turbidity, caused by run-terminating breakthrough or by
a coagulant feed malfunction, could be a signal for immediate corrective
action by the plant operator. Preventing the production of higher turbidity
water would prevent the sloughing of asbestos fibers from filters and would
thus improve filtered water quality.
Recommended Treatment Technique
On the basis of results presented in this report, granular media filters
can be operated to dependably remove asbestos fibers. The most important
115
-------
factor involved is properly conditioning the raw water before filtration so
that a very low effluent turbidity is produced. Alum and cationic polymers
have usually been used for coagulation. Two plants in Philadelphia success-
fully use ferric chloride however. Polymers often are used as coagulant aids
or filter aids. Control of coagulant dose, to assure thorough and complete
particle destabilization, is essential, as is careful control of process pH.
Filter rates from 2 to 10 gpm/sf have removed asbestos fibers, but abrupt rate
changes must be avoided, particularly if floe is not strong and could pass
through the filter as a result of a rate increase. An appropriate goal for
filtered water turbidity is 0.10 ntu, a goal that is now being met at some
water filtration plants. Asbestos fiber counts are likely to increase very
substantially as filtered water turbidity rises, so increases in turbidity
over 0.10 ntu should be interpreted as a signal for corrective action by
filter plant operators. Because turbidity is only an indicator of the
presence of particulates in water, attaining the lowest possible filtered
water turbidity is perhaps the best way to assure that a filtration plant is
operating at maximum effectiveness.
DIATOMITE FILTRATION FOR ASBESTOS FIBER REMOVAL
Amphibole Removal
Amphibole fibers were effectively removed by diatomaceous earth filtra-
tion at Duluth. In.their reviews of the Duluth pilot plant study, both
Baumann (37) and Logsdon and Symons (33) cited the use of alum-coated diato-
mite as particularly effective. Baumann suggested that alum-coated filter
media would have a positive surface potential, and that amphibole particles in
the pH range of 7 to 8 (Lake Superior raw water pH) would be negatively
charged. This charge difference would enhance the attraction of amphiboles to
the diatomite filter aid particles. Surface attachment could occur, facilita-
ting the removal of amphibole fibers too small to be removed by straining.
The use of both pressure and vacuum diatomite filters at Duluth permitted
comparison of the two kinds of equipment. Pressure filtration was more effec-
tive than vacuum filtration for removing amphiboles and turbidity-causing par-
ticulates. This may have been the result of raw water quality and a funda-
mental difference in the two types of filters. The pressure filter pump was
on the influent side of the filter. It provided a positive pressure, or
driving force through the filter, of up to 100 feet of water, as observed in
head loss measurements in some early long filter runs. Water passed through
the pressure filter and was discharged at atmospheric pressure.
In the vacuum filter operation, the driving force was the difference be-
tween atmospheric pressure at the influent side of the filter and the suction
at the pump on the effluent side of the filter. Head loss for the vacuum
filter runs was less than 15 feet of water in most runs.
When the clear, cold, oxygen-saturated water from Lake Superior passed
through the vacuum filter and pressure dropped below one atmosphere, the dis-
solved gases tended to come out of solution and form bubbles in the filter
cake. These bubbles could grow and collapse, permitting solids to bleed
through the filter cake. From June 21 to August 27, according to pilot plant
116
-------
operators' notes, bubbles were observed on the vacuum filter septum in 18 out
of 67 runs. Because of the air bubble problem associated with filtering very
cold waters, pressure filtration is better suited to such situations.
During the last half of the Duluth pilot plant study, pressure D.E. fil-
tration repeatedly produced water with turbidity of 0.10 ntu or lower. For
amphibole removal, the use of pressure filtration with alum coated diatoma-
ceous earth would be recommended. The grade of diatomite selected should be
fine enough to produce a filtered water of 0.10 ntu or lower, because high
fiber counts were usually associated with turbidities above 0.10 ntu (Fig. 33).
Nine of twelve BDL counts occurred with turbidities of 0.10 or lower.
If the septum pores are too large to permit effective precoating with the
fine diatomite, a two-stage precoating process could be used, with the coarser
DE coated on the septum and then a second layer of fine DE placed over the
first precoat layer. Only the second layer of precoat in such a case would
need to be alum-coated. Two-stage precoat was used often at Duluth.
Chrysotile Removal
Lawrence and coauthors have presented the best available information on
the use of diatomaceous earth filtration for removing chrysotile fibers from
water (21, 27). A reduction of 99.9 percent could be attained when waters
with 100 X 106 f/L to 1000 X 106 f/L were treated by filtraton through ordin-
ary diatomaceous earth. Hyflo-supercel* was used in their early work (21).
The type of diatomaceous earth used for later laboratory and field studies
was not identified.
Lawrence and Zimmerman (27) reported that the chrysotile fiber count in a
synthetic (laboratory-prepared) suspension could be reduced from 1000 X 10"
f/L to less than 0.1 X 10" f/L when it was filtered through diatomaceous earth
coated with aluminum hydroxide. This is more than a 99.99 percent reduction,
about the same degree of reduction that was attained by the filtration plant
at Duluth during storms on Lake Superior.
Field tests at Asbestos, Quebec, verified the efficacy of diatomite fil-
tration. Raw water contained 1000 X 10° f/L. Water filtered through uncoated
diatomite had 3 X 106 f/L, and water filtered through alum coated DE had 0.08
X 106 f/L, for a reduction exceeding 99.99 percent.
Alum coated diatomaceous earth is obviously more effective for chrysotile
than uncoated DE. Because both the coated diatomite and the individual chry-
sotile fibers are probably positively charged, the improved filter performance
would not be explained merely on the basis of surface potential. The Seattle
granular media filtration results, in which alum and cationic polymers removed
chrysotile may help explain this finding. Such removal would not be explained
on the basis of surface change alone, but it could be explained if chrysotile
fibers were associated with negatively charged particles, such as clays, and
were removed with them.
*Johns-Manville Products Corp., Denver, Colorado
117
-------
i.Or
0.3
\
>^
-------
Other Applicatons for D.E. Filtration
Diatomaceous earth filtration has been used for many years in industry to
filter both process flow streams and waste streams. Because of the advantages
of low capital cost and ease of operation, industries that need to filter
waste streams containing asbestos fibers are likely to adopt diatomite filtra-
tion. Both the Canadian research and the Duluth results suggest that this
process should be effective for treating industrial waste discharges and pot-
able waters.
COSTS
The cost of water treatment to attain water quality goals is an important
consideration in selecting a treatment technique. Cost aspects also may be
evaluated when quality goals are set; for example, the amount of hardness re-
moved at a softening plant. Because of the importance of costs, they will be
discussed in this portion of the report.
Treatment Costs Attributed to Asbestos
In Volume 2 of a report prepared for EPA, Gumerman, Gulp and Hansen dis-
cussed added treatment costs for asbestos fiber removal (50). They emphasized
the need to produce finished water turbidities of 0.1 to 0.2 ntu to accom-
plish virtually complete removal of asbestos fibers. Because consistently pro-
ducing high quality water would be necessary, the authors stated, ". . .a need
may arise for more highly skilled operators. Such additional cost would also
have to be charged to asbestos removal."
Gumerman et al. (50) listed a number of potential modifications to exist-
ing plants to enhance asbestos removal. These should also be considered for
new plants built for asbestos removal. They are:
1. Multi-stage rapid mix
2. Provision to feed polymers
3. Use of a pilot filter or coagulant control center
4. Use of mixed-media filters
5. Turbidity monitoring equipment for each filter
6. Higher than normal backwash rates
7. Washwater recovery
8. Lagoon or landfill disposal of sludge
9. Laboratory instrument to measure asbestos fiber removal
(a particle counter)
Some of the above items may have been incorporated into water filtration
plants constructed or modified in the 1960's or 1970's. Most were used at
Duluth and Two Harbors. An explanation of how to estimate the cost of the
above items is given by the authors in Volume 1 (50).
Capital Costs
Cost estimates for direct filtration plants by Gumerman et al. (50) were
compared with actual plant costs by Logsdon, Clark and Tate (51). Application
of direct filtration is promising when raw water is of such good quality that
119
-------
it previously was not filtered before being distributed to consumers. This is
the most likely situation in which discovery of asbestos in drinking water
would necessitate building a new filtration plant. Emphasis is placed on
direct filtration, therefore, in a discussion of plant costs.
Data on plant costs are available for Duluth, Two Harbors, and Silver Bay.
Duluth and Two Harbors have new filtration plants. The plant at Silver Bay
was upgraded, as described earlier. Costs for these plants, and a cost esti-
mate prepared by CH2M-Hill for a Seattle Tolt treatment plant, are given in
Table 45. The Minnesota cost data are updated to a May, 1979 Engineering
News-Record construction Cost Index of 2981. The Seattle ENR index of 3215
was used for the Tolt plant. In Table 45 a column for maximum plant capacity
normalized to a filtration rate of 5 gpm/sf is also shown. The rationale for
this, given by Logsdon, Clark, and Tate (51), was discussed in an earlier
paper by Dickson (52).
TABLE 45. TREATMENT PLANT CAPITAL COSTS
Treatment
Plant
Maximum
Capacity
mgd
Capacity
at 5 gpm/sf
mgd
Cost Year
million
$
Cost Adjusted
to May, 1979
million $
Silver Bay 2.3
(modification of
existing plant)
Two Harbors 2.6
Duluth 36
Seattle 100
(estimated)
*Seattle CCI used
4.4
3.2
30
72
1.0
1.5
6.2
1978
1977
1976
1.1
1.7
7.7
25*
When the costs for the four plants are updated and plotted on a capacity-
cost curve with the EPA estimate (50) also updated and shown, the specific
costs tend to be greater than the EPA estimates (Fig. 34). The differences in
costs may reflect a number of factors. Provision of sludge handling and
disposal at the plants was omitted in the EPA estimate, although it could
be calculated and included in the curve because CWC Engineers included
waste disposal cost curves in their report. Construction costs in Minnesota
may be higher than the national average because of winter conditions.
Construction costs are definitely higher than the national average in
Seattle, and the Seattle project cost estimate even included a $1.2 million
item for state sales tax. The Seattle conceptual design and the three
plants built on Lake Superior were intended specifically for asbestiform fiber
removal. Design engineers may have been more careful to provide process flexi-
bility and redundancy so that the plants would perform at a very high fiber
removal efficiency on a dependable basis.
120
-------
(-0
100
W
5 10
W
C
o
CO
O
O
E 1.0
<
o
0.1
Two Harbors^-
Silver Bay -
Seattle estimate-
Duluth
EPA estimate C = 0.51 Q°-66 (May 1979)
W Plants indexed to May 1979
10
CAPACITY, mgd (at 5 gpm/sf)
100
Figure 34. Direct filtration plant capacity vs. cost.
-------
Operating Costs
Plant operating costs vary over a wide range, so the extent to which
asbestos removal would add to the costs is not readily defined. Certainly
additional costs would occur at a plant that in normal operating practice was
barely meeting the 1 ntu limit. At a plant already producing filtered water
at the 0.1 ntu level, extra operating costs would be much lower.
Examples of actual costs for operation and maintenance for a number of
plants were given in the paper on costs of direct filtration. Those data are
shown in Table 46 and shown in Figure 35 which shows the downward trend in
costs as plant size increases. Costs were updated from July, 1977 to May,
1979 using the Wholesale Price Index ratio of 239/196.
TABLE 46. WATER FILTRATION PLANT OPERATION AND MAINTENANCE COSTS
Water Production O&M Costs
Utility and Average Maximum Average/Maximum Indexed to 5/79
Location MGD MGD percent ^/IQOO gallons
Clackamas Water Dist., 5.3 20 26.5 5.7
Clackamas, Oregon
Duluth, Minnesota 17 36 47.2 5.5
Bellingham, Washington 9.8 36 27.2 7.2
Metropolitan Water 19 54 35.2 5.1
Board, Oswego, N.Y.
East Bay MUD 10* 60 16.7 7.2*
Oakland, Calif.
Walnut Greek Plant
Springfield, Mass. 26 60 43.3 3.2
Southern Nevada Water 67 200 33.5 3.7
System, Las Vegas '
*EBMUD data reflect California drought conditions and water conservation.
At a more typical 12 to 13 mgd, filtration O&M costs would have been close
to 6.2 (6/1000 gallons.
The cost of treatment chemicals would be one of the concerns of plant
operators and managers who would need to filter water to remove asbestos
fibers. Chemical prices vary according to type and strength, quantity pur-
chased, arid shipping costs. No prices are given here. However, actual chemi-
cal doses used at some treatment plants or pilot plants removing asbestos
fibers are given so treatment plant personnel can formulate judgments on the
nature of the chemical costs they might experience. These are shown in Table
47. Chemical doses are low for the very clear waters and certainly not un-
reasonably high for the turbid waters. In fact, the chemical costs for the
122
-------
CO
30
(0
c
CO
O)
g 10
O
o
00
O
0.1
EPA estimate for 40% load factor
O FIELD DATA
10
WATER PRODUCED, mgd
100
Figure 35. Operation and maintenance cost vs. water production
for direct filtration. (May 1979)
-------
plants cited are probably a lot lower than chemical costs at a number of lime
softening plants.
Extensive cost data are not yet available because plants have not been
operated specifically for asbestos fiber removal for very long at this time.
As more years of operating experience are acquired, defining costs more close-
ly will be easier.
Monitoring Costs
Although capital and operating costs for filtration plants intended to
remove asbestos fibers might not be much greater than costs normally associ-
ated with filtration plants, the cost of monitoring could be an overwhelming
financial burden. In 1979 the only method that yields actual asbestos fiber
counts in water samples is the transmission electron microscope method. The
annual cost of submitting only two samples per week to a laboratory for analy-
sis would be in the range of $30,000 to $40,000. Only the largest utilities
can absorb such a high analytical cost.
Even if TEM analysis could be afforded, the results often are returned to
the water utility from the analyst four to six weeks after a sample was sub-
mitted. Therefore, electron microscopy is suitable only for providing histor-
ical records of water quality.
Day-to-day and minute-to-minute quality monitoring must be done by use of
turbidimeters. For most effective monitoring of filtered water turbidity, a
continuous recording turbidimeter with an alarm should be installed on each
filter, or a valving and timer arrangement should be installed so that one con-
tinuous turbidimeter could monitor three or four filters once each hour on a
rotating basis.
Economic Burden of Filtration for Asbestos Removal
Adopting new or improved treatment technology for contaminant removal
generally results in higher water treatment costs. Water utilities that build
filtration plants to remove asbestos fibers will incur significant capital and
operating costs. Utilities that have filtration plants that need adjustments
to operating procedures in order to attain greater reduction of the fiber
counts will also face increased expenses, but these will be much lower than
the costs of new plant construction.
In this century many water utilities have demonstrated that filtration
is affordable by building, operating, and maintaining water filtration plants.
A number of water utilities that filter surface waters meet the AWWA Quality
Goal of 0.1 ntu for filtered water, thus showing that attaining this degree of
clarity in filtered water is economically feasible. The degree to which water
utilities go to remove asbestos fibers from drinking water may depend more
upon their perception of the need for asbestos removal than on the economic
factors involved.
124
-------
TABLE 47. CHEMICAL DOSES FOR ASBESTOS FIBER REMOVAL
Treatment Plant Raw Water
Source Turbidity, ntu
Lakewood(Duluth) L. Superior
Two Harbors L. Superior
Seattle pilot plant Tolt Reservoir
Bollman plant Contra Costa Canal
CCCWD or Sacramento R.
Torresdale plant Delaware R.
Queen Lane plant Schuylkill R. and
Wissohickon Creek
Belmont plant Schuylkill R.
1 (10-15 max)
1 (10-15 max)
1 (5 max)
20-40
5-10
(max > 50)
5-10
(max > 170)
5-10
(max > 400)
Treatment Chemical
Alum
Nonionic Polymer
Sodium Hydroxide
Alum
Nonionic Polymer
Sodium Hydroxide
Alum
Lime
Nonionic Polymer
or
Alum
Cationic Polymer
Nonionic Polymer
or
Cationic Polymer
Alum
Polymer
Lime
Sodium Hydroxide
Ferric Chloride
Lime
Ferric Chloride
Lime
Alum
Lime
Dose
mg/L
10-15
1.5-4
0.07-012
15
0.4
11-12
7-10
1-4
0.02-0.25
3-5
2
0.1-0.3
3
40-55
0.02-0.03
3-6
9-15
about 10
about 10
about 20
-------
REFERENCES
1. Cunningham, H.M. , and R. Pontefract. Asbestos Fibres in Beverages and
Drinking Water. Nature, 232(5309): 332-333, 1971.
2. Cook, P.M., G.E. Glass, and J.H. Tucker. Asbestiform Amphibole Minerals:
Detection and Measurement of High Concentrations in Municipal Water
Supplies. Science, 185(4154): 853-855, 1974.
3. McCabe, L. J. and J. R. Millette. Health Effects and Prevalence of
Asbestos Fibers in Drinking Water. Presented at the 1979 Annual Confer-
ence of the American Water Works Association, San Francisco, California,
June 24-29, 1979. 18 pp.
4. Gross, P. et. al. Ingested Mineral Fibers: Do They Penetrate Tissue or
Cause Cancer? Archives of Environmental Health, 29(6):341-347, 1974.
5. Pontefract, R.D., and H.M. Cunningham. Penetration of Asbestos Through
the Digestive Tract of Rats. Nature, 243(5406): 352-353, 1973.
6. Cook, P.M., and G.F. Olson. Ingested Mineral Fibers: Elimination in
Human Urine. Science, 204(4389): 195-198, 1979.
7. Hallenbeck, W.H., and K. Patel-Mandlik. Fate of Ingested Chrysotile
Fiber in the Newborn Baboon. EPA-600/1-78-069, U.S. Environmental Pro-
tection Agency, Cincinnati, Ohio, 1978. 16 pp.
8. Meylan, W.M., et. al. Chemical Market Input/Output Analysis of Selected
Chemical Substances to Assess Sources of Environmental Contamination:
Task III. Asbestos. EPA-560/6-78-005, U.S. Environmental Protection
Agency, Washington, D.C., 1978, 314 pages.
9. Olson, H.L. Asbestos in Potable Water Supplies. Journal American Water
Works Association, 66(9): 515-518, 1974.
10. Buelow, R. W., et. al. The Behavior of Asbestos-Cement Pipe Under Vari-
ous Water Quality Conditions: A Progress Report, Part 1 - Experimental
Results. Presented at 1979 Annual Conference of the American Water Works
Association, San Francisco, California, June 27, 1979, 56 pp.
11. Millette, J.R., P- J. Clark, and M. F. Pansing. Exposure to Asbestos
from Drinking Water in the United States. EPA-600/1-79-028, U.S. Envir-
onmental Protection Agency, Cincinnati, Ohio, 1979, 87 pages.
126
-------
12. Black & Veatch, Consulting Engineers. Direct Filtration of Lake Super-
ior Water for Asbestiform Fiber Removal: Project Report and Appendices
A through I. EPA-670/2-75-050a through EPA-670/2-75-050g, U.S. Environ-
mental Protection Agency, Cincinnati, Ohio, 1975.
13. Logsdon, G. S., F. X. Schleppenbach, and T. M. Zaudtke. Filtration Works
Out Asbestos Fibers. Water & Sewage Works, 126(10):44-46, 1979.
14. Kay, G. Ontario Intensifies Search for Asbestos in Drinking Water.
Water and Pollution Control, 111(9): 33-35, 1973.
15. Kay, G. H. Asbestos in Drinking Water. Journal American Water Works
Association, 66(9): 513-514, 1974.
16. Kay, G. H. Letter of May 23, 1974 to 0. J. Schmidt, Black & Veatch Con-
sulting Engineers, Kansas City, Missouri.
17. Wigle, D. T. Cancer Mortality in Relation to Asbestos in Municipal Water
Supplies. Archives of Environmental Health, 32(4): 185-190, 1977.
18. Sargent, H. E., Asbestos in Drinking Water. Journal New England Water
Works Association, 88(1): 44-57, 1974.
19. Kanarek, M. S. Asbestos in Drinking Water and Cancer Incidence. Ph.D
Dissertation, University of California, Berkeley, California, 1978.
374 pp.
20. McMillan, L. M., R. G. Stout, and B. F. Willey. Asbestos in Raw and
Treated Water: An Electron Microscopy Study. Environmental Science &
Technology, 11(4): 390-394, 1977.
21. Lawrence, J. H. M. Tosine, H. W. Zimmerman, and T.W.S. Pang. Removal of
Asbestos Fibres from Potable Water by Coagulation and Filtration. Water
Research, 9(4): 397-400, 1975.
22. Lawrence, J. and H. W. Zimmerman. Potable Water Treatment for Some
Asbestiform Minerals: Optimization and Turbidity Data. Water Research,
10(3): 195-198, 1976.
23. Yao, K. M., M. T. Habibian, and C. R. O'Melia. Water and Wastewater Fil-
tration: Concepts and Applications. Environmental Science & Technology,
5(11): 1105-1112, 1971.
24. O'Melia, C. R. The Role of Polyelectrolytes in Filtration Processes.
EPA-670/2-74-032, U.S. Environmental Protection Agency, Cincinnati, Ohio,
1974. 82 pp.
25. Letterman, R. D. A Study of the Treatment of Lake Michigan Water Using
Direct Filtration. UILU-WRC75-102, University of Illinois Water Re-
sources Center, Urbana, Illinois, 1975. 102 pp.
127
-------
26. Trussell, R.R. Application of Treatment Technology. Prepared for U.S.
EPA Environmental Research Center Technology Transfer Seminar on Design-
ing and Upgrading Drinking Water Systems. Portland, Oregon, May 25-26,
1977.
27. Lawrence, J., and H. W. Zimmerman. Asbestos in Water: Mining and Pro-
cessing Effluent Treatment. Journal Water Pollution Control Federation,
49(1): 156-160, 1977.
28. Hunsinger, R. B., J. Lawrence, and K. J. Roberts. Chrysotile Asbestos
Dilution Studies. Report #66, Ontario Ministry of the Environment and
Canada Centre for Inland Waters, Toronto and Burlington, Ontario, 1977.
10 pp.
29. Hunsinger, R.B., J. Lawrence, and K. J. Roberts. Pilot Plant Studies to
Effect Chrysotile Asbestos Fibre Reduction During Potable Water Treat-
ment. Report #67, Ontario Ministry of the Environment and Canada Centre
for Inland Waters, Toronto and Burlington, Ontario, 1977. 36 pp.
30. Foley, P. D. and G. A. Missingham. Monitoring Public Water Supplies.
Journal American Water Works Association, 68(2): 105-111, 1976.
31. Watkins, J. Jr., R. A. Ryder, and W. A. Persich. Investigation of Tur-
bidity, Asbestos Fibers and Particle Counting Techniques as Indices of
Treatability of -a Cascade Mountain Water Source. In: Proceedings of the
American Water Works Association Annual Conference, Atlantic City, New
Jersey, 1978. Part 2, pp. 333 and following.
32. Schmitt, R. P., D. C. Lindsten, and T. F. Shannon. Decontaminating Lake
Superior of Asbestos Fibers. Environmental Science & Technology, 11(5):
462-465, 1977.
33. Logsdon, G. S., and J. M. Symons. Removal of Asbestiform Fibers by Water
Filtration. Journal American Water Works Association, 69(9):499-506, 1977.
34. The Johns-Manville Corporation. Johns-Manville Comments on "Interim
Report on Water Supply for the Duluth-Cloquet-Superior Area." Denver,
Colorado, 1975. 36 pp.
35. Black & Veatch, Consulting Engineers. Duluth-Superior Urban Study:
Interim Report on Water Supply for the Duluth-Cloquet-Superior Area,
Technical Report, Appendix 2. U.S. Army Corps of Engineers, St. Paul
District, March, 1975. 482 pp.
36. Robinson, J. H. et. al. Direct Filtration of Lake Superior Water for
Asbestiform Solids Removal. Journal American Water Works Association,
68(10): 531-539, 1976.
37. Baumann, E. R. Diatomite Filters for Asbestiform Fiber Removal from Water.
Presented at American Water Works Association Annual Conference, Minneap-
olis, Minn., 1975. (printed as Appendix I of EPA/670/2-75-050, cited as
Reference 12).
128
-------
38. Pattern, J. L. Unusual Water Treatment Plant Licks Asbestos Fiber Prob-
lem. Water & Wastes Engineering, 14(11):41-44 and 50, 1977.
39. Peterson, D. L. The Duluth Experience: Asbestos, Water and the Public.
Journal American Water Works Association, 70(1): 24-28, 1978.
40. Logsdon, G.S. Direct Filtration - Past, Present, Future. Civil Engineer-
ing-ASCE, 48(7):68-73. 1978.
41. Kirmeyer, G. J., G. S. Logsdon, J. E. Courchene, and R. R. Jones. Re-
moval of Naturally Occurring Asbestos Fibers from Seattle's Cascade
Mountain Water Source. Presented at American Water Works Association
Annual Conference, San Francisco, California, 1979.
42. Kirmeyer, G. J. Seattle Tolt Water Supply Mixed Asbestiform Removal
Study (project report and two appendix reports). EPA-600/2-79-125,
EPA-600/2-79-126, and EPA-600/2-79-153; U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1979.
43. American Public Health Association, American Water Works Association, and
Water Pollution Control Federation. Standard Methods for the Examination
of Water and Wastewater, Fourteenth Edition. APHA, Washington, D.C.,
1975.
44. Anderson, C. H. and J. M. Long. Fibrous Asbestos, Preliminary Interim
Procedure, Transmission Electron Microscopy. Unpublished EPA document,
1976. This has been superceded by Anderson, C. H., and J. M. Long.
Interim Method for Asbestos in Water. U. S. Environmental Protection
Agency, Athens, Georgia, 1978 (to be published),,
45. Stone, B. G. Design of Ralph D. Bollman Water Treatment Plant. Journal
American Water Works Association, 64(6): 369-376, 1972.
46. Harris, W. L. High-Rate Filter Efficiency. Journal American Water Works
Association, 62(8): 515-519, 1970.
47. Research Built This Water Plant. The American City, 80(7): 82-85, 1965.
48. Jardine, J. W. Chicago's Filtration Plant a Winner! Water Works and
Wastes Engineering, 2(5): 47-51, 1965.
49. Diehl, S. R., D. T. Smith and M. Sydor. Optical Detection of Fiber Par-
ticles in Water. EPA-600/2-79-127, U.S. Environmental Protection Agency,
Cincinnati, Ohio, 1979. 71 pp.
50. Gumerman, R. C., R. L. Gulp, and S. P. Hansen, Estimating Costs for
Water Treatment as a Function of System Size and Treatment Efficiency,
EPA-600/2-78-182, U.S. Environmental Protection Agency, Cincinnati, Ohio,
1978.
129
-------
51. Logsdon, G. S., R. M. Clark & C. H. Tate. Costs of Direct Filtration to
Meet Drinking Water Regulations. To be published in Journal American
Water Works Association.
52. Dickson, R. D. Estimating Water System Costs, In: Water Treatment Plant
Design, R. Sanks, editor. Ann Arbor Science Publishers, Inc., Ann Arbor,
Michigan, 1978. pp. 763-799.
130
-------
APPENDIX A-l. DULUTH MONITORING DATA FROM LAKEWOOD FILTRATION PLANT
Date
1977
1/1
1/11
1/13
1/18
1/18
1/20
1/21
1/22
1/23
1/24
1/25
1/26
1/27
1/28
1/29
1/30
2/01
2/03
2/05
2/12
2/17
2/19
3/02
Plant Influent
Turbidity Amphibole
(ntu) (106 fibers/liter)
-
-
0.43
0.43
0.41
0.52
0.39
0.54
0.45
0.36
0.33
0.30
0.38
0.37
0.35
0.30
0.36
0.28
0.28
0.26
0.44
0.37
_
-
-
37
41
31
40
43
44
44
35
38
19
29
35
42
29
49
41
55
20
17
-
Plant Effluent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.09
0.05
0.07
0.07
0.06
0.10
0.05
0.04
0.05
0.07
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.05
0.06
0.40
0.30
0.62
BDL(0.19)
0.60
0.26
BDL(0.19)
0.20
0.10
BDL(0.19)
0.067
BDL(0.067)
0.13
BDL(0.067)
11
"
"
131
-------
APPENDIX A-l. (Continued)
Date
1977
3/09
3/21
4/04
4/11
4/18
4/25
5/02
5/10
5/16
6/02
6/14
7/01
7/08
7/13
7/18
7/25
8/01
8/08
8/15
8/24
8/30
Plant Influent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.47
0.41
0.30 38
0.30
0.30
0.30
0.30
0.50
0.50
0.40
0.62 134
1.6
0.45
0.56 145
0.60 130
0.33
0.55
0.60
0.80
0.82 75
0.40 24
Plant Effluent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.04
0.04
0.04
0.05
0.04
0.04
0.04
0.06
0.07
0.06
0.05
0.08
0.042
-
0.04
0.04
0.04
0.04
0.035
0.035
0.04
0.096
BDL(0.067)
»
BDL(0.067)
0.091
0.036
0.030
0.050
0.070
0.064
0.064
0.26
BDL(0.064)
-
0.067
0.078
0.038
0.019
0.077
0.086
0.058
132
-------
APPENDIX A-l. (Continued)
Date
9/06
9/14
9/19
9/19
9/28
10/04
10/08
10/12
10/14
10/19
10/25
10/31
11/07
11/16
11/22
12/02
12/07
12/14
12/20
12/28
Plant
Turbidity
(ntu) (1C
0.68
0.68
10.0
14.0
1.3
2.6
11.0
2.3
-
0.85
0.63
0.56
0.51
0.47
0.62
0.48
0.83
0.52
1.4
0.48
Influent
Amphibole
1200
69
72
830
120
-
26
18
20
5.6
34
6.6
6.5
90
3.8
120
27
Plant
Turbidity
(ntu) (10
0.035
0.035
0.045
0.04
0.035
0.045
-
0.033
0.035
-
-
0.038
0.039
0.031
0.038
0.034
0.037
0.035
0.036
Effluent
Amphibole
6 fibers/liter)
0.038
0.019
0.048
BDL( 0.010)
0.010
0.029
-
0.053
0.019
0.010
0.010
0.048
0.088
0.022
0.038
0.010
0.019
0.010
BDL(O.OIO)
133
-------
APPENDIX A-l. (Continued)
Date
1978
1/04
1/10
1/18
1/24
2/06
2/14
2/24
2/27
3/06
3/16
3/21
3/28
4/05
4/11
4/17
4/25
5/1
5/8
5/15
5/23
5/30
6/5
6/12
Plant Influent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.46
0.47
0.42
0.40
0.86
0.56
0.57
-
0.51
0.48
0.72
0.68
0.82
1.5
0.73
0.89
0.94
0.62
0.52
0.49
15.0
1.0
0.80
60
21
24
27
75
6.1
44
-
55
48
40
36
100
250
-
210
200
160
110
160
500
130
170
Plant
Turbidity
(ntu) (10
0.036
0.038
0.035
0.038
0.038
0.034
-
0.034
0.038
0.039
0.039
0.037
0.039
0.051
0.038
0.041
0.033
0.034
0.033
0.034
0.039
0.035
0.034
Effluent
Amphibole
6 fibers/liter)
0.029
0.010
0.010
0.014
0.019
0.010
-
0.067
BDL(O.OIO)
0.019
0.019
0.010
0.029
0.019
0.058
BDL(0.019)
BDL(0.019)
0.019
0.019
0.048
0.029
0.058
0.010
134
-------
APPENDIX A-l. (Continued)
Date
iy78
6/19
6/26
7/5
7/11
7/17
7/25
8/02
8/09
8/17
8/22
9/06
9/13
9/22
9/28
10/5
10/12
10/20
10/27
11/3
11/10
11/22
12/08
12/15
Plant Influent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.72
0.81
0.77
0.67
1.05
0.68
0.64
0.61
0.75
0.35
0.71
2.4
0.84
2.0
0.67
0.37
0.38
0.73
0.65
7.7
0.74
-
-
180
99
27
96
45
46
59
2.2
36
27
12
160
15
13
16
33
16
150
68
460
6.5
-
-
Plant
Turbidity
(ntu) (10
0.041
0.040
0.040
0.035
0.040
0.041
0,032
0.031
0.039
0.032
0.041
0.04
0.04
0.052
0.003
0.04
0.031
0.048
0.031
0.04
0.032
0.031
0.065
Effluent
Amphibole
6 fibers/liter)
0.019
0.019
0.010
BDL(O.OIO)
0.038
-
0.058
0.048
0.048
0.019
0.019
0.0096
0.0096
BDL(0.0096)
0.029
BDL(0.0096)
BDL(0.0096)
0.010
BDL(O.OIO)
0.029
0.023
BDL(O.OIO)
BDL(O.OIO)
135
-------
APPENDIX A-l. (Continued)
Date
1979
1/5
1/16
1/25
1/30
2/5
2/13
2/19
2/27
3/8
3/9
3/14
3/20
3/22
4/3
4/5
4/12
4/17
4/24
5/4
5/10
5/17
5/18
5/18
Plant Influent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.43
-
-
-
0.65
-
-
0.21
-
0.19
0.16
-
0.39
-
0.24
0.35
0.26
1.3
0.44
8.4
2.2
-
-
34
-
-
-
81
-
-
18
-
19
14
-
24
-
5.4
11
9.6
40
70
96
47
-
-
Plant Effluent
Turbidity Amphibole
(ntu) (106 fibers/liter)
0.039
0.033
0.033
0.048
0.048
0.036
0.031
0.033
0.033
-
0.043
0.040
-
0.033
-
0.038
0.040
0.039
0.039
0.036
0.037
0.038
0.038
BDL(O.OIO)
BDL(O.OIO)
BDL( 0.010)
BDL(O.OIO)
0.019
BDL(O.OIO)
BDL(O.OIO)
BDL(O.OIO)
0.010
-
BDL( 0.010)
BDL(O.OIO)
-
0.010
-
0.010
BDL(O.OIO)
0.010
0.019
0.019
0.019
0.0096
0.0096
136
-------
APPENDIX A-l. (Continued)
Date
1979
5/22
5/23
5/29
5/31
6/4
6/4
6/6
6/11
6/11
6/12
6/19
6/21
6/25
Plant Influent Plant Effluent
Turbidity Amphibole Turbidity Amphibole
(ntu) (106 fibers/liter) (ntu) (106 fibers/liter)
1.2 46 0.040
0.039
0.052
0.038
0.048
1.3 56 0.043
0.038
0.043
1.5 96 0.044
0.039
0.045
0.040
1.8 150 0.043
0.041
0.051
0.041
1.3 180 0.033
0.031
0.033
0.033
1.6 170 0.036
0.032
0.010
0.038
0.019
0.048
0.077
0.029
0.086
0.067
A
A
A
0.029
0.010
BDL( 0.010)
0.12
0.019
0.048
0.029
0.019
0.12
0.31
0.010
*These samples are in the process of being analyzed.
137
-------
APPENDIX A-2. EFFECTS OF FILTRATION RATE CHANGES ON WATER QUALITY AT DULUTH
Filtered Water Quality
OJ
00
Date
HL at
rate
change
ft water
Media Equilibrium during
Normal Operation
Amphibole
ntu 106 f/L
Highest during
Rate Change
Amphibole
ntu 106 f/L
Approximate
Average for
30 minutes
During Rate
Change
Amphibole
106 f/L
Unit Processes Employed
Before Filtration
Rapid Floccu- Sedimen-
Mix lation tation
Ra'te Increased From 3.25 to 4.33 gpm/sf When One Filter Was Shut Down for Backwash
5/24/77 2.0 dual
6/1/77
6/9/77
2.0 dual
7.2 dual
0.09
0.06
0.10
9/28/77 7.5 dual
10/4/77 2.0 dual
0.04
0.04
5/25/77 2.1 mixed 0.06
6/2/77 2.3 mixed 0.06
6/14/77 6.7 mixed 0.05
10/14/77 4.5 mixed 0.034
2/27/78 2.8 mixed 0.07
0.07
0.06
0.06
0.04
0.04
0.06
0.06
0.06
0.05
0.07
0.16 0.38
0.07 0.38
0.44 12.8
0.08 0.06
0.05 0.30
0.06
0.06
0.06
0.04
0.04
0.32
0.096
0.13
0.05
0.07
0.2
0.2
3-4
0.04
0.1
0.2
0.1
0.1
0.05
0.05
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
APPENDIX A-2. EFFECTS OF FILTRATION RATE CHANGES ON WATER QUALITY AT DULUTH
Filtered Water
Quality
Date HL at Media Equilibrium during Highest during Approximate Unit Processes Employed
rate Normal Operation Rate Change Average for Before Filtration
change Amphibole Amphibole 30 minutes Rapid Floccu- Sedimen-
ft water ntu 106 f/L ntu 106 f/L During Rate Mix lation tation
Change
Amphibole
106 f/L
7/6/77 2.0
9/20/77 2.5
11/11/77 3.5
11/28/77 9.0
8/10/77 2.0
8/17/77 9.0
11/22/77 2.7
11/28/77 6.9
Rate Increased From 0
dual 0.04 0.06
dual 0.05 0.02
dual 0.04 0.07
dual 0.07 0.03
mixed 0.07 0.05
mixed 0.05 0.08
mixed 0.04 0.02
mixed 0.04 0.03
to 3.25
0.05
0.05
0.04
0.71
0.05
0.12
0.04
0.04
gpm/sf When Entire Plant Started Up
0.48 0.2 X X
0.19 0.05 X X
0.09 0.05 X
3.7 1 X
0.11 0.1 X X
2.2 1 XX
0.12 0.05 X
0.04 0.05 X
X
X
X
X
-------
APPENDIX A-2. EFFECTS OF FILTRATION RATE CHANGES ON WATER QUALITY AT DULUTH
Filtered Water Quality
Date HL at Media Equilibrium during Highest during Approximate Unit Processes Employed
rate Normal Operation Rate Change Average for Before Filtration
change Amphibole Amphibole 30 minutes Rapid Floccu- Sedimen-
ft water ntu 106 f/L ntu 106 f/L During Rate Mix lation tation
Change
Amphibole
106 f/L
Rate Increased From 0 to 3.25 gpm/sf in One Filter When Operation Resumed After Backwash
6/22/77 2.0 dual 0.05 0.03 0.10 2.3 2 X
3/6/78 2.5 dual 0.04 0.01 0.04 0.25 0.1 XXX
11/23/77 2.0 mixed 0.04 0.03 0.05 0.10 0.05 X
3/10/78 2.4 mixed 0.04 0.02 0.04 0.21 0.1 XXX
-------
APPENDIX B. SUMMARY OF SEATTLE PILOT PLANT ASBESTOS DATA
Run #
3-R
3-F
4c-R
4c-F
5c-R
5c-F
6-R
6-F
11-R
11-F
12d-R
12d-F
21-R
21-F
21-F
21-F
21-F
21-F
24-R
24-F
24-F
24-F
29-R
29-F
29-F
Hour
Into
Run
8
8
5
5
5
5
4
4
6
6
7
7
7
2
6
7
8
12
7
6
7
13
10
10
17
Turbidity
(NTU)
0.1
1.4
0.1
1.4
0.08
1.3
0.07
1.15
0.28
1.0
0.09
0.66
0.065
0.06
0.34
0.07
0.059
0.60
0.085
0.36
0.062
0.62
0.090
0.10
Raw (fibers/liter)
Amphibole
(106)
5.7
5.7
3.31
3.31
3.06
3.06
3.46
3.46
4.33
4.33
1.76
1.76
2.18
2.18
2.18
2.18
2.18
2.18
2.4
2.4
2.4
2.4
0.94
0.94
0.94
Chrysotile
(in6)
8.9
8.9
5.12
5.12
16.39
16.39
13.0
13.0
13.29
13.29
13.14
13.14
25.8
25.8
25.8
25.8
25.8
25.8
9.4
9.4
9.4
9.4
4.25
4.25
4.25
Finished (fibers/liter)
Amphibole
(106)
0.04(NSS)
0.05 (NSS)
<0.01(ND)
O.OS(NSS)
0.42
__
O.Ol(NSS)
0.01 (NSS)
O.Ol(NSS)
0.72(NSS)
<0.01(ND)
<0.01(ND)
0.04(NSS)
0.6 (NSS)
0.04(NSS)
<0.01(ND)
cO.Ol (ND)
Chrysotile
(106)
._
0.09
0.09
0.15
0.15
1.64
__
0.13
0.16
0.16(NSS)
12.25
0.19
0.09
0.34
6.2
0.13
0.07
0.22
Removal
Amphibole
(%)
99.4
98.5
>99.6
98.6
90.3
__
99.4
99.5
99.5
67.0
>99.5
>99.5
98.3
75.0
98.3
>98.9
>98.9
Chrysotile
(%)
99.0
98.2
99.1
98.8
87.7
__
99.0
99.4
99.4
52.5
99.3
99.7
96.4
72.3
98.6
98.4
94.8
(continued)
141
-------
APPENDIX B (CONTINUED)
Run #
33-R
33-F
33-F
44-R
44-F
44-F
51-R
51-F
51-F
53-R
53-F
53-F
62-R
62-F
62-F
62-F
70-R
70-F
70-F
70-F
89-R
89-F
89-MM
89-MM
89-MM
Hour
Into
Run
6
6
9
11
11
13
9
3
3
9
3
9
16
3
9
16
15
5
11
15
9
9
9
15
19
Turbidity
(NTU)
0.61
0.062
0.053
0.56
0.042
0.09
0.54
0.07
0.07
0.50
0.13
0.24
0.35
0.10
0.105
0.37
0.35
0.085
0.072
0.08
0.35
0.064
0.06
0.049
0.065
Raw (fibers/liter)
Amphibole
(106)
0.65
0.65
0.65
0.90
0.90
0.90
<0. 29 (ND)
ND
ND
0.70(NSS)
0.70
0.70
<0.12(ND)
ND
liD
ND
<0.07 (ND)
ND
ND
ND
<0.05(ND)
ND
ND
ND
ND
Chrysotile
(10<5)
3.82
3.82
3.82
2.8
2.8
2.8
8.4
8.4
8.4
3.6
3.6
3.6
2.52
2.52
2.52
2.52
2.81
2.81
2.81
2.81
1.2
1.2
1.2
1.2
1.2
Finished (fibers/liter
Amphibole
(106)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(MD)
<0.01(ND)
<0.02 (ND)
1.2 (HSS)
<0.02 (ND)
<0.02 (ND)
0.1 (NSS)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
Chrysotile
(106)
0. 06 (NSS)
0.22
0.26
<0.01(ND)
0.06 (NSS)
0.20
--
0.27
11.2
0.34
0.24
2.28
--
0.06(NSS)
0.26
0.17
0.014(NSS)
0.043(NSS)
0.04 (NSS)
0.17
Removal
Amphibole
(%)
>98.4
>98.4
>98.8
>98.8
--
"
Chrysotile
(%)
98.4
94.2
--
90.7
>99.6
99.3
97.6
92.5
+211.1
86.5
90. 5
9.5
97.9
90.7
94.0
98.8
96.4
96.7
85.8
(continued)
142
-------
APPENDIX B (CONTINUED)
Run #
93-R
93-MM
93-MM
93-MM
93-MM
93-F
108-R
108-MM
108-MM
108-MM
108-MM
108-F
111-R
111-MM
111-MM
111-MM
111-MM
111-MM
111-MM
111-F
120-R
120-R
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
120-MM
Hour
Into
Run
18
3
12
14
17
18
4
2
10
15
16
18
12
7
12
16
17
18
19
19
9
15
1
2
6
9
12
13
15
16
17
18
19
20
Turbidity
(NTU)
0.38
0.098
0.10
0.065
0.074
0.08
0.55
0.12
0.073
0.081
0.068
0.082
0.85
0.004
0.19
0.082
0.09
0.10
0.22
0.072
3.3
3.4
0.14
0.096
0.08
0.065
0.06
0.07
0.062
0.071
0.071
0.115
0.14
0.28
Raw (fibers/liter)
Amphibole
(106)
<0.07 (ND)
ND
ND
ND
ND
ND
<0.14(ND)
ND
ND
ND
ND
ND
<0. 14 (ND)
ND
ND
ND
ND
HD
ND
ND
0.19(NSS)
<0.10 (ND)
0.19(NSS)
0.19U1SS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19(NSS)
0.19 (NSS)
Chrysotile
(106)
3.6
3,6
3.6
3.6
3.6
3.6
3.61
3.61
3.61
3.61
3.61
3.61
4.62
4.62
4.62
4.62
4.62
4.62
4.62
4.62
5.38
10.1
5.38
5.38
5.38
5.38
5.38
5.38
5.38
5. 38
5.38
5.38
5.33
5.38
Finished (fibers/liter)
Amphibole
(106)
<0.02(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01 (ND)
--
<0.02(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.02 (ND)
<0. 01 (ND)
<0.01 (ND)
<0.02(ND)
<0.02 (ND)
<0.01 (ND)
<0.02 (ND)
<0.02 (ND)
.01(ND)
<0.01 (ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01 (ND)
<0.01 (ND)
<0.02(ND)
<0.04 (ND)
Chrysotile
(106)
0.09(NSS)
0.53
0.07
0.18
0.04(NSS)
0.34
0.03(NSS)
0.14
0.07 (NSS)
0.13
0.09
0. 07 (NSS)
0.07 (NSS)
0.03 (NSS)
0.07 (NSS)
0.15
0.04 (NSS)
0.14
0.19
0.07
<0.01(ND)
O.l(NSS)
<0.01 (ND)
0.01 (NSS)
0.01 (NSS)
0.02 (NSS)
<0.01 (ND)
0.19
0.14 (NSS)
Removal
Amphibole
(%)
_
~
Chrysotile
(%)
97.5
85.3
98.0
98.9
98.9
90.6
99.2
96.1
98.1
96.4
98.1
98.5
98.5
99.4
98.5
96.8
99.1
97.4
96.5
98.7
>99.8
99.8
>99.8
99.8
99.8
99.6
>99.8
96.5
97.4
(continued)
143
-------
APPENDIX B (CONTINUED)
Run #
120-MM
120-MM
120-MM
135-R
135-MM
135-MM
135-MM
135-MM
135-MM
135-MM
135-MM
135-MM
135-MM
135-MM
135-MM
161-R
161-CC
161-CC
161-CC
161-CC
161-CC
161-FC
161-FC
161-FC
161-FC
161-FC
161-MM
161-MM
161-MM
161-MM
161-MM
174-R
174-MM
174-MM
174-MM
Hour
Into
21
22
23
9
0
1
2
3
4
11
13
21
22
23
24
6
1
4
10
13
15
1
4
10
13
15
1
4
10
13
15
1.5
0
1
1.5
Turbidity
(NTU)
0.48
0.57
1.2
1.8
0.28
0.14
0.10
0.096
0.090
0.074
0.11
0.065
0.089
0.072
0.065
0.34
0.092
0.085
0.069
0.082
0.082
0.095
0.071
0.069
0.072
0.073
0.135
0.082
0.079
0.089
0.081
0.36
0.089
0.075
0.20
Raw (fibers/liter)
Amphibole
(106)
0.19(NSS)
Q.19(NSS)
0.19(NSS)
<0.07 (ND)
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
<0.04(ND)
ND
TSID
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
0.07 (NSS)
Chrysotile
(10<5>
5.38
5.38
5.38
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
1.84
"
Finished (fibers/liter)
Amphibole
(106)
<0.04(ND)
<0.07 (ND)
<0.08(ND)
<0.02(ND)
<0.02(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01 (ND)
<0.01 (ND)
<0.01(ND)
<0.'01(ND)
<0.01 (ND)
<0.01(ND)
<0.01(ND)
<0.01 (ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01 (ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.03 (ND)
Chrysotile
(106)
0.28
0.57
1.25
0.31
0.32
0.14
0.2
0.08
0.04(NSS)
O.OS(NSS)
0.02 (NSS)
<0.01(ND)
O.Ol(NSS)
0.01 (NSS)
<0.01(ND)
<0.01(ND)
0.01 (NSS)
<0.01(ND)
<0.01(ND)
<0.01(ND)
0.04(NSS)
0.02 (NSS)
0.03(NSS)
<;0.01(ND)
0. 01 (NSS)
0.02(NSS)
<0.01(ND)
<0. 01 (ND)
<0.01(ND)
0.07
0.1
0.36
Removal
Amphibole
(%)
"*
^^
Chrysotile
(%)
94.8
89.4
76.8
--
92.1
91.8
96.4
94.9
97.9
99.0
98.7
99.5
>99.7
99.7
99.7
--
>99.5
>99.5
>99.5
99.5
>99.5
>99.5
98.0
98.0
99.0
98.5
>99.5
99.5
99.0
>99.5
>99.5
96.2
94.5
80.4
(continued)
144
-------
APPENDIX B (CONTINUED)
Run #
174-MM
174-MM
174-MM
174-MM
174-MM
174-MM
174-MM
174-MM
174-CMM
L74-CMM
174-CMM
174-CMM
174-CMM
174-CMM
174-CMM
174-CMM
174-CMM
174-CMM
174-CRM
174-CMM
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
174-CC
Hour
Into
Run
2
3
4
5
6
7
14
18
0
1
1. 5
2
3
4
5
6
7
8
14
18
0
1
1.5
2
3
4
5
6
7
8
14
19
Turbidity
(NTU)
0.079
0.078
0.075
0.079
0.072
0.075
0.078
0.075
0.095
0.070
0.21
0.068
0.070
0.071
0.062
0.089
0.075
0.070
0.065
0.079
0.097
0.07
0.20
0.070
0.071
0.090
0.071
0.081
0.068
0.070
0.070
0.070
Raw (fibers/liter)
Amphibole
(106)
__
Chrysotile
(106)
__
--
__
Finished (fibers/liter)
Amphibole
(106)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01 (ND)
<0.01 (ND)
<0.01 (ND)
<0.01 (ND)
<0.02 (I1D)
<0.01(ND)
<0.01 (ND)
<0.01(ND)
<0.01 (ND)
<0.01(ND)
<0.01 (ND)
^O.Ol(ND)
<0.01(ND)
<0.01 (ND)
<0.01(KD)
<0.01(ND)
<0.03 (ND)
<0.01(ND)
<0. 01 (ND)
<0.01 (ND)
<0.01 (ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
<0.01(ND)
Chrysotile
(106)
0.03(NSS)
<0.01 (ND)
O.Ol(NSS)
O.Ol(NSS)
O.Ol(NSS)
0.02(NSS)
<0.01 (ND)
<0.01(HD)
0.09
0.02 (NSS)
0.94
<0.01(ND)
0.03(NSS)
0.01 (NSS)
0.06(USS)
0.02(NSS)
0.01 (NSS)
0.01 (NSS)
0.01 (NSS)
0.05 (NSS)
O.Ol(NSS)
0.04(NSS)
<0.03(ND)
0.01 (NSS)
<0. 01 (ND)
<0.01(ND)
<0.01(ND)
<0. 01 (ND)
0. 06 (NSS)
<0.01(ND)
<0.01 (ND)
O.Ol(ND)
Removal
Amphibole
(%)
~~
_
_-
~~
Chrysotile
(%)
98.4
>99.4
99.4
99.4
99.4
98.9
>99.4
>99.4
95.1
98.9
48.9
>99.4
98.4
99.4
96.7
98.9
99.4
99.4
99.4
97.3
99.4
97.8
> 98.3
99.4
>99.4
>99.4
>99.4
>99.4
96.7
>99.4
>99.4
>99.4
NOTES:
R = Raw Sample
F Finished Sample from Waterboy
MM = Finished Sample from Mixed Media Filter Column with MS-6 Sand
CMM = Finished Sample from Mixed Media Filter Column with MS-18 Sand
FC = Finished Sample from Dual Media Filter Column with Fine Coal
CC = Finished Sample from Dual Media Filter Column with Coarse Coal
ND = None Detected
NSS = Not Statistically Significant (4 or fewer fibers actually counted)
145
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-
>06
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
WATER FILTRATION FOR ASBESTOS FIBER REMOVAL
5. REPORT DATE
December 1979 (Issuing DateL
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Gary S. Logsdon
8. PERFORMING ORGANIZATION REPO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Same as below
10. PROGRAM ELEMENT NO.
1CC614
11. CONTRACT/GRANT NO.
Inhouse
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory--Cin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Summary 5/74 - 6/79
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Contact: Gary S. Logsdon (513) 684-7345
16. ABSTRACT
This report presents a comprehensive review of data on removal of asbestos
fibers by granular media filtration and diatomaceous earth filtration. It summarizes
data obtained in pilot plant studies at Duluth and Seattle, in research program carried
out at Duluth's Lakewood filtration plant, and monitoring at Silver Bay and Two Harbors,
Minnesota plants, Chicago, Philadelphia, and in the San Francisco Bay area.
Chrysotile and amphibole fiber concentrations in drinking water can be substantiall
reduced by granular media filtration. Reductions of up to 99.99 percent were reported
during storm conditions at Duluth, Minnesota. Effective granular media filtration re-
quired careful control of pH, coagulant doses, and filtered water turbidity.
Research to date indicates that coating the diatomaceous earth filter aid with alum
num hydroxide substantially increases the removal of both amphibole and chrysotile fiber
Duluth results indicate that filtered water turbidity should be 0.10 ntu for most
effective fiber removal.
When a granular media filtration plant is properly operated, turbidity readings can
be an indicator of fiber removal efficiency even though turbidity cannot directly
measure asbestos fibers in the concentrations found at water treatment plants. Filtered
water turbidity should be 0.10 nephelometric turbidity units (ntu) or lower to maximize
fiber removal. Turbidity increases of 0.1 or 0.2 ntu above this value generally were
accompanied by large increases in fiber concentrations.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDEDTERMS
c. COSATI Field/Group
Asbestos, filtration, potable water,
turbidity, water treatment, amphiboles
Seattle, Lake Superior,
San Francisco Bay area,
Philadelphia, Chicago,
fiber removal, chrysotile
diatomaceous earth filter
granular media filter
13 B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
160
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
146
i U S GOVERNMENT PRINTING OFFICE 1981
10-657-146/5500
------- |