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<
-------
                               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 not•experience 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 data—Analysis 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 data—Both 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—
     Monitoring—The 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 storms—During 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 systems—Unfiltered 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 quality—When 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 mixers—Several 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 mixers—Several 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 systems—Both 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.

     Flocculator—To 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



* *%

. **
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  data—Kirmeyer  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 water—The 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 give•additional 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 data—The 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 data—One 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 results—Kirmeyer (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 mixing—Rapid 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.

      Flocculation—The  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).

     Sedimentation—Sedimentation 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 control—The 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 dose—In 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.

     Media—The 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 control—During 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 filters—In 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-WRC—75-102, 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. 33—3 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

-------