See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/250142660

Treatment selection guidelines for particle and

NOM removal

Article in Aqua · September 2009

DOI: 10.2166/aqua.2009.201

CITATIONS READS

5 92

3 authors, including:

Matthew Valade James Edzwald

Hazen and Sawyer, P.C. Clarkson University
10 PUBLICATIONS 81 CITATIONS 92 PUBLICATIONS 3,646 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by James Edzwald on 15 August 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 Provided for non-commercial research and educational use only.
Not for reproduction or distribution or commercial use.

This article was originally published by IWA Publishing. IWA Publishing recognizes
the retention of the right by the author(s) to photocopy or make single electronic
copies of the paper for their own personal use, including for their own classroom use,
or the personal use of colleagues, provided the copies are not offered for sale and
are not distributed in a systematic way outside of their employing institution.

Please note that you are not permitted to post the IWA Publishing PDF version of
your paper on your own website or your institution’s website or repository.

Please direct any queries regarding use or permissions to [email protected]

 424 Practical Paper Q IWA Publishing 2009 Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

Practical Paper

Treatment selection guidelines for particle

and NOM removal
M. T. Valade, W. C. Becker and J. K. Edzwald

ABSTRACT

Proper selection of appropriate treatment processes for surface water supplies is a function M. T. Valade (corresponding author)
Hazen and Sawyer, P.C.,
of raw water quality including particles and natural organic matter (NOM). This paper presents 24 Federal Street, Suite 302,
Boston MA 02110,
an updated foundation for process selection based on routinely measured parameters of particles
USA
(turbidity) and organic matter content (UV254, color and TOC). The distinction between mineral Tel.: +1 617 574 4747
Fax: +1 617 574 4799
and non-mineral turbidity is addressed. Processes considered include direct filtration, E-mail: [email protected]

conventional sedimentation, dissolved air flotation and contact clarification. The results of this W. C. Becker
Hazen and Sawyer, P.C.,
paper are based on a survey of currently operating facilities, pilot studies, theory and North 498 Seventh Avenue, 11th Floor,
American experience. Recommended guidelines for process selection as a function of raw water New York NY 10018,
USA
TOC, color and turbidity are presented. Figures to aid process selection based on source water
J. K. Edzwald
quality are provided. Department of Civil and Environmental
Engineering,
Key words | filtration, flotation, NOM, particles, sedimentation, treatment selection University of Massachusetts.
Amherst MA 01003,
USA

INTRODUCTION

Achieving finished drinking water quality that meets supply is important to be made early in the planning process
standards has typically been feasible through conventional as it can save time and reduce project costs. Furthermore,
treatment of coagulation, flocculation, sedimentation, gran- selection of the most appropriate process is critical to
ular media filtration and disinfection for a wide range of raw achieving consistent high quality finished water. Poor
water qualities. New standards and goals that focus on process selection can cause a utility to waste significant
reducing disinfection by-products and on reducing exposure capital and operating costs. Therefore, a practical approach
to pathogens in drinking water are requiring water utilities to process selection is warranted.
to provide improved treatment. Many utilities will be Janssens & Buekens (1993) presented an approach
required to upgrade existing processes in order to meet based on raw water turbidity and chlorophyll a for
these new requirements and will need to select treatment preliminary assessment and selection of appropriate
processes that will provide optimum treatment based on processes for removal of particles and algae as shown in
their raw water quality parameters. Figure 1. Sedimentation, dissolved air flotation and single-
Proper selection of appropriate treatment processes for stage direct filtration, all widely used treatment processes,
surface water supplies is a function of raw water quality were included in their analysis. The less common process
including particles and organic content. Identification of of two-stage direct filtration was also included in their
appropriate processes for treatment of a specific water assessment. The recommendations they set forth in Figure 1
doi: 10.2166/aqua.2009.201
 425 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

flocculation is not provided (termed contact or in-line

filtration). When flocculation is used, the detention time is
typically 10 –15 min compared to 30 min for a conventional
sedimentation plant. In direct filtration plants particle
removal occurs only in the filters, which typically consist
of dual media with filter loading rates of 10– 15 m/h. In
some installations deep bed mono- or dual-media
filters (anthracite) or filter adsorbers (e.g. granular activated
carbon (GAC)) have been used also at filter rates of
10– 20 m/h.

Conventional and high rate sedimentation

Figure 1 | Floc separation processes: selection diagram (Reprinted from Janssens &
Buekens 1993 J. Water Supply Res. Technol. AQUA 42(5), 279–288, with A conventional sedimentation plant is one that consists of
permission from the copyright holders, IWA Publishing).
coagulant addition followed by flocculation, large rectangu-
lar or circular sedimentation basins, and rapid media filters.
for treatment process selection were based on the authors’
Note that high rate plate and tube sedimentation processes
experience and their scientific and operational knowledge
are also considered in this paper. Coagulant is needed to
of drinking water treatment. Janssens & Buekens’ paper
destabilize particles and to provide metal hydroxide floc for
was an important contribution that aided process selec-
adsorption or co-precipitation of NOM. Since new particles
tion. However, the figure and their guidelines have some
are formed by precipitation, chemical coagulation dosing
limitations. First, the only measured parameter for natural
and pH conditions must be carried out such that these
organic matter (NOM) is chlorophyll a for algae. This only
particles are destabilized.
accounts for particulate NOM (algae), which for many
Flocculation (slow mixing) is needed to promote
water supplies is low compared to the dissolved NOM.
particle growth through particle –particle contacts. The
Also, it is not a routinely measured parameter for most
goal of flocculation in a conventional treatment plant is to
water utilities. Second, they do not distinguish mineral
make “settleable” floc, i.e. floc that will readily be removed
turbidity (clays, silts) from non-mineral turbidity (e.g.
in a settling basin. The surface loading (or overflow) rate for
algae). Finally, their guidelines do not consider total
conventional settling is in the range of 0.5– 1.0 m/h,
organic carbon (TOC).
depending on the type of settling process, nature of the
This paper presents an updated foundation for process
raw water turbidity (mineral or non-mineral) and water
selection based on more routinely measured parameters of
temperatures. Addition of plates or tubes to sedimentation
particles (turbidity) and organic matter content (UV254,
basins will reduce the footprint of the treatment process by
color and TOC). The distinction between mineral and
achieving an effective footprint settling rate of 2.5– 6.0 m/h.
non-mineral turbidity is addressed. Plant types considered
Granular media filtration follows sedimentation to remove
include direct filtration, conventional sedimentation, dis-
any remaining floc particles. Filters typically consist of
solved air flotation (DAF) and contact clarification. A brief
mono-media (sand) or dual-media (anthracite or GAC and
description of each process is provided below.
sand) with loading rates of 10 –20 m/h.

Direct filtration Dissolved air flotation

Direct filtration does not contain a clarification process A dissolved air flotation plant is similar to a sedimentation
but does require coagulant addition to destabilize particles plant except flocculated particles are separated from the
and to remove a small amount of TOC. In some cases, liquid stream by floating the floc to the water surface.
 426 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

Coagulant addition and flocculation are still used, but the primary sources were supplemented with data obtained
goal of flocculation in this case is to produce a floc that can be from utility databases maintained by equipment manufac-
removed by attachment to microscopic air bubbles. Floccu- turers for utilities that have recently integrated new
lation times for DAF facilities built prior to the mid-1990s treatment schemes (e.g. DAF, ballasted-sand, etc.) into
were 20 –30 min. Since then, it was demonstrated that their treatment plants, as well as from pilot studies related
shorter flocculation times are feasible (for example, Edzwald to upgrading existing facilities with new process schemes.
& Wingler 1990; Valade et al. 1996). Full-scale facilities are The Water:\Stats database included data for the
now often designed with flocculation times of 10 min or less following parameters: turbidity, TOC, DOC, color, UV254,
(for example, 5 min for the Croton Water Treatment Plant pH, alkalinity and hardness. Utilities contacted directly
for New York City (Crossley & Valade 2006; Crossley et al. were requested to provide average and maximum water
2007)). Two-stage flocculation is common. quality data for the previous three years for these same
The typical design surface loading rate of the DAF parameters. While the additional parameters beyond tur-
process usually varies between 10– 15 m/h—significantly bidity represent a more limited set of data compared to
higher than the loading rate on a conventional sedimen- turbidity, because fewer utilities make regular measure-
tation basin and higher than the footprint loading rates ments of these parameters, they are essential to determining
for high rate plate and tube sedimentation processes. the proper treatment process selection. Additionally, the
In addition, high-rate DAF processes have recently been utilities were queried regarding the type of clarification
developed at rates of 20– 40 m/h (Morris & Hess 2004; process utilized, the age of the treatment plant, whether any
Edzwald 2007). oxidant is used, the average and maximum coagulant
dosages over the three year period, the trophic state of the
source water and whether algae are a recurrent problem.
Contact clarification
These latter data are not included in Water:\Stats.
In this paper, contact clarification is a category used for a Data were grouped by plant type into four categories:
variety of processes that includes sludge-blanket clarifica- settling, DAF, direct filtration and contact clarification. As
tion, ballasted-sand clarification and contact adsorption noted above, contact clarification includes sludge-blanket
clarification. Although each of these specific processes is clarification, ballasted-sand clarification and upflow-fil-
different from a mechanical perspective, they all work by tration through plastic media (sometimes called adsorption
enhancing particle removal via particle – particle or par- clarification).
ticle– media contacts. All source water data were evaluated for average
and maximum turbidity conditions against average and
maximum organic content (i.e. TOC), as well as surrogate
organic measurements of UV254 and true color. The data in
METHODS
Figures 2 and 3 show that settling plants are used to treat a
As mentioned above, the selection of the proper treatment wide range of water quality—from supplies that are low in
plant is a function of raw water quality, finished water goals, turbidity and TOC to those containing high levels of
and coagulant dose and type. turbidity and/or TOC. Conventional settling was often the
Raw water quality data and process treatment infor- only clarification process considered by utilities in the USA
mation were collected from about 400 water treatment prior to the 1990s; hence, most plants built in the USA
plants across the United States and Canada. Primarily, data before 1990 utilize conventional sedimentation. However,
were obtained directly from contacting water utilities and the authors of this paper believe that, although sedimen-
from the Water:\Stats Water Utility Database (AWWA tation can be effectively utilized for a wide range of raw
1996), an in-depth database of water utility information water qualities, it is not necessarily the best treatment
compiled by the American Water Works Association process for all water qualities. This premise is expanded on
(AWWA) based on nationwide utility surveys. These below. Plots that focus on direct filtration, DAF and solids
 427 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

RESULTS AND DISCUSSION

For over a hundred years, sedimentation has been used

widely as the primary means for clarification of drinking
water with a broad range of source water qualities. In the
last 40 years, other plant types have been used for particular
source water qualities. Table 1 shows the range of ages, as
well as the average age, for the treatment plants that are
included in this analysis. We were able to obtain the ages of
about 25% of the 400 treatment plants: however, we believe
the distribution of data in Table 1 is representative and
Figure 2 | Survey source water average turbidity and TOC data. shows clearly that sedimentation has been historically the
predominant plant type. Direct filtration and contact
contact clarification were developed. The figures were clarification plants were introduced on a large scale starting

evaluated to ascertain the limits of the raw water quality in the 1960s and DAF plants are an even more recent

parameters for which the different treatment processes have development in the United States and Canada. The first

been used in practice. Based on the assumption that DAF plant in the United States using European-based
technology was placed into service in 1993 (Nickols et al.
sedimentation has been used historically for a wide range
1995). Note that the age of the direct filtration plants in this
of water qualities and potentially misapplied to some water
survey range from 0 to 72 years—aside from one direct
qualities, data for conventional settling were not included
filtration plant currently under construction in Canada, no
in plots provided hereafter in order to provide clarity. It
direct filtration plants were built in North America since the
should be noted that some data points for a specific
advent of DAF in the USA in 1993. Direct filtration plants
treatment type appeared to be outliers. In these cases, the
are limited to treating source waters with very low TOC
authors contacted the water utility to verify the effectiveness
concentrations (due to disinfection by-product concerns)
of the treatment process under the extreme water quality
and they lack a clarification process as an additional barrier
event. For cases when a utility reported the treatment
to pathogens. The plant currently being constructed in
process had difficulty achieving the required finished water
Vancouver, Canada includes UV disinfection as an
quality goals at the desired plant flow rate, the data were
additional treatment barrier for Cryptosporidium. Further-
removed from the figures.
more, the improved treatment efficiency and lower cost of
membrane filtration, particularly for waters with low
turbidity and NOM where pretreatment is not required,
will likely mean few new direct filtration plants will be
constructed in the USA and Canada.

Table 1 | Summary of water treatment plant ages

Water treatment plant age

Plant type Average (yr) Range (yr) No. of plants in data

Settling 45 2 – 107 39
Direct filtration 29 0 – 72 13
Dissolved air flotation 4 0 – 14 33
Solids contact clarifiers 35 11– 74 7

Figure 3 | Survey source water maximum turbidity and TOC data. Note that plants with age of 0 are currently under construction.
 428 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

criteria for direct filtration: average turbidity ,5 ntu and

average TOC ,3 mg/L.
The early DAF plants in the USA and Canada were built
largely to treat high quality source waters. In many cases the
supplies were previously unfiltered and DAF was chosen
over direct filtration. This biased the survey data since we
know from experience around the world that DAF plants
have been built to treat waters with a range of raw water
qualities, especially when it comes to TOC and non-mineral
turbidity (supplies with algae). The survey data show that
DAF plants typically have average raw water turbidity levels
less than 10 ntu and average TOC levels up to 10 mg/L;
Figure 4 | Survey average source water turbidity and TOC data from Figure 2 however, there is no reason why DAF plants would not be
excluding settling plant data set.
suitable for supplies with higher TOC and non-mineral

Turbidity and TOC turbidity levels. High TOC supplies would require higher
coagulant doses, but coagulant– NOM flocs have low
The collected data were analyzed in detail for each plant densities which makes them suitable for removal by
type—direct filtration, conventional clarification, DAF and DAF. Likewise, non-mineral turbidity also has a low
solids contact clarification. A plot of average source water density which would be removed by DAF even at high
turbidity versus average TOC for each plant type is concentrations.
presented in Figure 4. The data show that direct filtration The data for solids contact clarifiers included in
plants are used on the highest quality sources—those with
Figure 2 span the , 1 –70 ntu turbidity range and have
average turbidity values generally less than 3 ntu and
average TOC levels up to 6 mg/L. Note that for clarity in
average TOC levels less than 3.5 mg/L. This makes sense
Figures 4 – 9, the data in the high turbidity range have not
because all of the particle removal in a direct filtration plant
been shown.
occurs in the filters. Higher turbidity or higher TOC levels
A plot of maximum turbidity vs. maximum TOC levels is
would result in higher coagulant doses, causing excessive
presented in Figure 5. The data show a similar trend to that
filter headloss development and short filter runs. Higher
noted above—direct filtration plants are used to treat the
TOC waters containing even a modest fraction of humic
highest quality source water, followed by DAF and then
matter demand higher coagulant doses (Pernitsky &
Edzwald 2006).
Janssens & Buekens (1993) recommended direct fil-
tration as a feasible treatment process for waters with
turbidities less than 10 ntu and 10 mg/L chlorophyll a—see
Figure 1. Chlorophyll a is not addressed in this paper;
instead the focus is on TOC (a more readily measured
parameter as discussed above). The TOC data in Figure 4
show an upper limit for direct filtration plants of 3.5 mg/L
TOC. Pernitsky & Edzwald (2006) suggested direct filtration
is feasible for source waters with TOC , 3 mg/L, which is
close to this upper limit in the survey data.
Based on the survey data, the Janssens and Buekens’
criteria, the Pernitsky and Edzwald criteria, and our own
Figure 5 | Survey source water maximum turbidity and TOC data from Figure 3
experience, we recommend the following source water excluding settling plant data set.
 429 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

solids contact clarification. The maximum turbidity for a

solids contact clarifier was 2,000 ntu (not shown). The
maximum turbidity value for DAF was 100 ntu. Although
the DAF process can operate at this raw water turbidity
level for some time, it would require higher recycle rates for
sufficient bubbles to lower floc density or, if typical recycle
rates of about 10% are used, then the clarified turbidity
would increase, causing higher particle loading on the filters
and more frequent backwashing. Although DAF is typically
not suited for high mineral turbidity levels, if the turbidity is
caused by organic constituents (e.g. algae), then DAF may
be a suitable process as mentioned above. Similarly, the
Figure 7 | Survey maximum source water turbidity and color.
maximum turbidity level for direct filtration in this survey
was approximately 90 ntu, which is also not sustainable for Inspection of the data for maximum color (Figure 7)
an extended period of time. Note that maximum turbidity reveals that direct filtration has been used to treat water
data for settling plants ranged up to 2,700 ntu (not shown in with color that reach up to 50 cu and DAF has been used to
Figure 5, shown in Figure 3). treat waters with color that exceeds 500 cu, although at
present for the USA and Canada most source waters with
DAF do not exceed 100 cu.
Turbidity and color Contact clarifiers have been used to treat a wider range

Average source water turbidity versus average color data are of water with an average color of typically between

plotted in Figure 6. All color data reported herein are true 20– 60 cu and maximum color of nearly 400 cu.

color. Not surprisingly, the color data reveal similar findings

to the turbidity and TOC data, with an upper turbidity limit Turbidity and UV254
of 3 ntu and 10 ntu for direct filtration and DAF, respect-
Average and maximum UV254 and turbidity values are
ively. Direct filtration plants have been constructed to treat
plotted in Figures 8 and 9, respectively. The data set is
water with average color up to 20 cu and DAF plants to
limited as fewer treatment plants reported taking UV254
treat water up to 200 cu, but again the authors believe there
measurements. (Only one contact clarification plant
is no upper limit for DAF plants as the coagulant –NOM
reported UV254 measurements and has not been included
flocs would be readily removed through flotation.
herein.) However, trends for direct filtration and DAF can

Figure 6 | Survey average source water turbidity and color. Figure 8 | Survey source water average turbidity and UV254.
 430 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

Figure 9 | Survey source water maximum turbidity and UV254.

Figure 11 | Process selection diagram based on maximum water quality.

be seen with direct filtration plants having source water

with average and maximum UV254 less than 0.1 cm21 and
1. Direct filtration plants are most suitable for relatively
0.15 cm21, respectively. Pernitsky & Edzwald (2006)
stable supplies with low turbidity and low organic matter
suggested that direct filtration be used only for water with
concentrations. High levels of turbidity or natural
UV254 , 0.07 cm21. DAF plants surveyed have source
organic matter (and associated increases in coagulant
waters with UV254 values up to 0.5 cm21 and 0.7 cm21 for
dose) will overload a direct filtration plant. If direct
average and maximum values, respectively.
filtration is determined to be an acceptable process by
the users, we suggest that it be limited to supplies with
average levels of turbidity ,5 ntu, color ,20 cu,
RECOMMENDED GUIDELINES TOC , 3 mg/L and UV254 , 0.07 cm21.
Recommended guidelines for process selection as a function 2. Dissolved air flotation plants are more robust than direct
of raw water TOC, color and turbidity are presented in filtration plants and can easily handle low to moderate
Figures 10 and 11. These guidelines are based on the survey turbidity supplies with high levels of organic matter and
data, the authors’ experience, and the Pernitsky & Edzwald color. The authors suggest that DAF plants be limited to
(2006) paper for direct filtration. UV254 is not included in supplies with average raw water mineral turbidity levels
the figures, but is included in the following text. Presen- of 10 ntu or less. Treatment of source waters with higher
tation and discussion follow: turbidities is feasible when the turbidity is caused by
organic constituents such as algae. There is no upper
boundary for TOC or color.
3. The performance of solids contact clarifiers is a
function of both raw water quality and the specific
type of clarifiers. In general, solids contact clarifiers
perform well over a wider turbidity range than direct
filtration and DAF. They may be particularly well
suited to supplies with a variable source water quality
(e.g. low average turbidity and organics, with high
maximum turbidity). Some solids contact clarifiers can
be problematic on supplies with rapidly changing
temperature or when rapid changes in plant flow are
Figure 10 | Process selection diagram based on average water quality. required.
 431 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

4. Conventional settling plants can handle the highest raw upper limit of organic content. Maximum turbidities can
water turbidity levels. They are most suited for supplies range up to 50 ntu (mineral sources) and 200 ntu (non-
with average raw water turbidities greater than 10 ntu. mineral sources).
Where conventional settling is used on supplies with † Solids contact clarifiers are recommended for water
low turbidity and low TOC, an increase in metal salt qualities having significant variations between average
coagulant dose is sometimes needed to make enough and maximum turbidities.
floc for the clarification process to be effective. When † Conventional settling can be used to treat a wide range of
polyelectrolytes or low metal salt coagulant doses are used raw water qualities, but are recommended for waters
in a conventional plant, the process tends to perform in a with maximum turbidities for mineral and non-mineral
direct filtration mode. sources of greater than 50 and 200 ntu, respectively.

CONCLUSIONS ACKNOWLEDGEMENTS
The paper by Janssens & Buekens (1993) provided treatment The authors would like to express their deep appreciation
plant selection guidelines with respect to raw water quality to all of the personnel at the water utilities who graciously
but was limited to raw water turbidity and chlorophyll a. provided information for this project and to the equipment
In this paper we extended the evaluation of raw water manufacturers—Infilco Degremont, Inc., ITT WWW Leopold
quality to include mineral and non-mineral turbidity and and Parkson Corporation—who provided data, as well as
various measures of organic matter such as TOC, true color contacts at many of the water utilities. Special thanks are
and UV254. Proper selection of treatment processes based extended to Mr Donald Brailey for the initial idea for this
on raw water quality characteristics including turbidity and paper, Mr Ian Crossley for his advice and assistance in
NOM is important. The recommended guidelines presented developing this paper, Ms Eileen Feldman for her efforts
in this paper provide a rational way for selecting appro- supporting the writing effort and to Ms Anni Luck for her
priate treatment processes based on raw water turbidity assistance in gathering the data and compiling it for analysis.
and NOM parameters such as color, TOC and UV254. Our
recommended guidelines are summarized in Figures 10 and
11. Conclusions for the selection of treatment plant type REFERENCES
based on the guidelines follow.
AWWA 1996 Water:\Stats Water Utility Database. AWWA,
† Direct filtration is recommended for high quality waters Denver, CO.
Crossley, I. & Valade, M. 2006 A review of the technological
with average raw water turbidities less than 5 ntu and
developments of dissolved air flotation. J. Water Supply Res.
maximum turbidities that do not exceed 30 ntu. Average Technol.—AQUA 55(7 –8), 479 –491.
raw water organic content should not exceed 3 mg/L of Crossley, I., Herzner, J., Bishop, S. & Smith, P. 2007 Going
TOC or 20 cu of color and maximum raw water organic underground— constructing New York City’s first water
treatment plant, a 1,100 Ml/d dissolved air flotation, filtration,
content of 5 mg/L of TOC or 35 cu of color. However,
and UV facility. In: Proc. 5th International Conference on
direct filtration may be further limited to waters with Flotation, Flotation 2007, Flotation in Water and Wastewater
even lower organic content if water age in the distri- Systems, Seoul, South Korea.
Edzwald, J. 2007 Developments of high rate dissolved air flotation
bution system leads to high levels of disinfection by-
for drinking water treatment. J. Water Supply Res.
products, thereby requiring a process that will effectively Technol.—AQUA 56(6 –7), 399 –410.
remove higher levels of organic matter. Edzwald, J. & Wingler, B. 1990 Chemical and physical aspects of
† DAF is recommended for relatively high quality waters dissolved air flotation for the removal of algae. J. Water Supply
Res. Technol.—AQUA 39, 24 – 35.
with average raw water turbidities below 10 ntu from
Janssens, J. & Buekens, A. 1993 Assessment of process selection for
river sources (i.e. mineral turbidity) or 100 ntu from particle removal in surface water treatment. J. Water Supply
reservoir sources (i.e. non-mineral turbidity), with no Res. Technol.—AQUA 42(5), 279 –288.
 432 M. T. Valade et al. | Treatment selection guidelines for particle and NOM removal Journal of Water Supply: Research and Technology—AQUA | 58.6 | 2009

Morris, S. & Hess, R. 2004 Next generation high rate dissolved air Pernitsky, D. & Edzwald, J. 2006 Selection of alum and
flotation (DAF): design and optimization of full-scale and pilot polyaluminum coagulants: principles and applications.
processes. In: WQTC Proceedings. American Water Works J. Water Supply Res. Technol.—AQUA 55(2),
Association, Boulder, CO. 121 –141.
Nickols, D., Moerschell, G. & Broder, M. 1995 The first DAF water Valade, M., Edzwald, J., Tobiason, J., Dahlquist, J., Hedberg, T. &
treatment plant in the United States. Water Sci. Technol. Amato, T. 1996 Pretreatment effects on particle removal by
31(3 –4), 239 –246. flotation and filtration. J. AWWA 88(12), 35 –47.

First received 23 December 2008; accepted in revised form 26 June 2009

View publication stats