Organic compounds and microbial assessment of a seawater reverse osmosis facility at Tampa Bay Water, USA

Abstract

The Tampa Bay Water seawater reverse osmosis (SWRO) facility is the first large-capacity seawater desalination plant in the United States. The feedwater source for the facility is an estuarine system that is biologically very productive and contains naturally-occurring high concentrations of algae, marine bacteria, total organic carbon (mostly dissolved), transparent exopolymer particles (TEP), the biopolymer fraction of natural organic matter, and phosphate. The high-organic composition of the feedwater places stress on the conventional sand pretreatment system utilized at the facility resulting in high organic passage into the membrane process and flow through into the permeate. In particular, the direct passage of particulate TEP (p-TEP) into the membranes has a major impact on the biofouling rate. Based on the data collected, the pretreatment is ineffective at removing key organic components that impact the rate of membrane biofouling, particularly bacteria and p-TEP. Perhaps the pretreatment could be re-designed to use a dissolved air floatation system (DAF) followed by ultrafiltration as a remedy that would likely move the biofouling problem to the ultrafiltration process, which has an easier cleaning process. Consideration could be given to using a groundwater source of feedwater as a permanent remedy to the operational issues.

Full text

Contents lists available at ScienceDirect
Desalination
journal homepage: www.elsevier.com/locate/desal
Organic compounds and microbial assessment of a seawater reverse osmosis
facility at Tampa Bay Water, USA
Natalie J. Harvey
a
, Zahid ur Rehman
b
, TorOve Leiknes
b
, Noreddine Ghaffour
b
,
Hidetoshi Urakawa
a,c
, Thomas M. Missimer
a,⁎
a
U. A. Whitaker College of Engineering, Emergent Technologies Institute, Florida Gulf Coast University, 16301 Innovation Lane, Fort Myers, FL 33913, United States
b
King Abdullah University of Science and Technology (KAUST), Water Desalination and Reuse Center (WDRC), Biological and Environmental Science and Engineering
Division (BESE), Thuwal, 23955-6900, Saudi Arabia
c
Department of Ecology and Environmental Studies, Florida Gulf Coast University, 10501 FGCU Boulevard South, Fort Myers, FL 33965-6565, United States
ARTICLE INFO
Keywords:
Tampa Bay Water
Seawater reverse osmosis (SWRO) desalination
Membrane biofouling
Total organic carbon (TOC)
Transparent exopolymer particles (TEP)
Marine bacteria
ABSTRACT
The Tampa Bay Water seawater reverse osmosis (SWRO) facility is the first large-capacity seawater desalination
plant in the United States. The feedwater source for the facility is an estuarine system that is biologically very
productive and contains naturally-occurring high concentrations of algae, marine bacteria, total organic carbon
(mostly dissolved), transparent exopolymer particles (TEP), the biopolymer fraction of natural organic matter,
and phosphate. The high-organic composition of the feedwater places stress on the conventional sand pre-
treatment system utilized at the facility resulting in high organic passage into the membrane process and flow
through into the permeate. In particular, the direct passage of particulate TEP (p-TEP) into the membranes has a
major impact on the biofouling rate. Based on the data collected, the pretreatment is ineffective at removing key
organic components that impact the rate of membrane biofouling, particularly bacteria and p-TEP. Perhaps the
pretreatment could be re-designed to use a dissolved air floatation system (DAF) followed by ultrafiltration as a
remedy that would likely move the biofouling problem to the ultrafiltration process, which has an easier
cleaning process. Consideration could be given to using a groundwater source of feedwater as a permanent
remedy to the operational issues.
1. Introduction
Shortages of freshwater throughout the world have driven the de-
velopment of cost-effective desalination technologies [1,2]. Desalina-
tion was performed in the past using mostly thermal distillation pro-
cesses until the commercialization of seawater reverse osmosis (SWRO)
membranes in the late 1950s [3]. Today, reverse osmosis (RO) is the
most economic desalination technique for producing freshwater from
seawater [3–5]. Despite the SWRO process being the most energy-effi-
cient desalination process, it is plagued by the issue of membrane
biofouling [6–11]. Extensive use of engineered pretreatment or use of
subsurface intake systems have been used to reduce the rate of bio-
fouling, but cannot eliminate it [12,13].
A wide variety of pretreatment schemes arranged in different
manners have been used to reduce the rate of membrane biofouling,
including debris removal, coarse and fine sand filtration, sand filtration
with added coagulation and flocculation, dissolved air flotation (DAF),
membrane filtration, and ultrafiltration (UF) [14,15]. Despite the use of
various pretreatment processes, Negati et al. [12] reported that 70% of
SWRO facilities in the Middle East have had biofouling problems that
have impacted operations and treatment costs.
The process of SWRO membrane biofouling is quite complex and
involves both the biochemistry of the raw feedwater [11], nutrient
concentrations [16–17], water temperature [18], and the SWRO plant
design and perhaps operation methods [19,20]. One measure of the
general biochemistry of the feedwater is the concentration of total or-
ganic carbon (TOC). A comparison of the TOC concentrations to the
number of annual cleanings of membranes required in SWRO facilities
is given in Table 1.
The data show that in the SWRO facilities using open-ocean intakes,
there is a general relationship between the concentration of TOC and
the required frequency of membrane cleanings, which is a proxy for the
rate of biofouling. It is understood that not only the TOC is a factor, but
the various forms of organic matter in the feedwater also influence the
rate of biofouling [11]. The very high rate of cleanings required in the
Arabian Gulf plant may be due to a combination of relatively high TOC
https://doi.org/10.1016/j.desal.2020.114735
Received 22 June 2020; Received in revised form 25 August 2020; Accepted 27 August 2020
⁎
Corresponding author.
E-mail address: tmissimer@fgcu.edu (T.M. Missimer).
Desalination 496 (2020) 114735
0011-9164/ © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T

values and high water temperatures [16–18]. The highest TOC values
were found at the Tampa Bay Water facility, which has experienced
difficulties from the onset of operation [21].
In 1998, Tampa Bay Water awarded a contract to Stone and Webster
and Poseidon Resources Corporation for the design and construction of
an SWRO plant. However, by 2000, Stone and Webster went bankrupt
and Poseidon took over full ownership of the project and partnered
with Covanta Tampa Construction. In December 2001, Covanta failed
in their financial obligations and Tampa Bay Water resumed ownership
of the project in March 2002, when the plant was only 50% complete
[23,24]. Covanta then failed to meet two construction-related mile-
stones, and litigation ensued wherein Covanta, which declared bank-
ruptcy in 2003, agreed to a settlement transferring the rights to operate
the plant to Tampa Bay Water. The plant experienced problems during
the initial startup where membranes lasted months instead of years and
cartridge filters that should last months lasted only weeks [23–25]. In
2003, the plant encountered severe difficulties with the pretreatment
system wherein the membranes began to foul excessively and rapidly.
Many changes were made to the plant with the expectation that the life
of the RO membranes would be extended, including coating filters with
diatomaceous earth, so that fine size particles were extracted by the
sand filters [24]. However, the plant was still unable to properly op-
erate at full capacity due to membrane fouling and expense issues [26].
In 2004, Tampa Bay Water contracted with American Water-Acciona
Agua to resolve the design and construction deficiencies and complete a
redesign for the plant in 2007 [27].
The redesign included modifying the plant to improve pretreatment
efficiency by correcting inadequate screening and changing the filtra-
tion design. Changes were also made in the post-treatment process to
meet the capacity goal of 94,697 m
3
/day (25 MGD) of drinking water in
2010 [26,28]. Specifically, precoat filtration was installed to augment
the existing sand filters, removing particles too fine to be extracted by
the sand filters, which then extended the life of the RO membranes. The
plant initially had two-stage sand filtration, which was then converted
to a single-stage upward-flow system, thereby doubling the number of
first stage sand filters so that individual filters could operate at lower
feed rates and have greater efficiency. New piping and controls for the
influent water to be dispersed evenly were also implemented. For the
intake pumps, variable-frequency drives were added to help maintain a
constant water temperature and stabilize water chemistry. Improved
screening and chemical feed systems were then added to the intake
pipeline to inhibit biological growth. The SWRO membrane system was
redesigned with added pipes, pumps, and electrical systems to allow for
an easier cleaning process that was more reliable. Lastly, lime from a
new stem to stabilize the treated water was added to provide additional
protection to the distribution pipelines [24].
While considerable improvements have been made to increase the
productivity of the Tampa Bay Water SWRO plant, there are still on-
going operational challenges. The plant is operated intermittently to
meet seasonal peak demands within the overall Tampa Bay Water
service area. In the 2019 water year, it operated from December 2018
to May 2019. In the 2020 water year, it will operate from December to
mid-July. However, the SWRO membranes must be cleaned every two
months, which affects the plant availability, membrane lifetime, as well
as costs and environmental issues. The environmental issue of concern
is the disposal of a greater amount of turbid water that must be treated
before discharge.
It is the purpose of this research to make a comprehensive assess-
ment of the organic carbon passing from the feedwater through each
process within the plant in time, including RO permeate, to assess the
causes of membrane biofouling and to make comparisons to the op-
eration to other SWRO facilities. Water samples were collected during
each month of operation during the 2019 dry season. This is a parti-
cularly important investigation because this plant uses feedwater that is
exceptionally high in organic carbon concentrations compared to other
global facilities of high capacity. An important question is why the
Table 1
Comparison of SWRO plants showing TOC concentration versus the number of cleanings per year. Note that the Arabian Gulf plant changed membranes from polyamide to cellulose triacetate because of the high
requirement of cleaning. The Jeddah plant has used cellulose triacetate (CTA) membranes from the beginning of operations. CTA membranes are chlorine tolerant [22]. At the Arabian Gulf plant, the inflow water is
chlorinated for 1 h every 24 h and requires membrane cleaning 2 times per year. At the Red Sea facility, the raw water is chlorinated for 1 h every 8 h and the membranes require cleaning 3 to 4 times per year. Note that
the pretreatment methods are given for the original SWRO plant design for the large plants on the Arabian Gulf and the Red Sea. It should be noted that the change in the membrane type from polyamide to CTA did not
prevent biofouling, but it did reduce the rate. No pretreatment process is going to totally stop biofouling, but the issue is to cut the rate, which reduces the frequency of membrane cleaning.
Site TOC of seawater (mg/L) Intake type Pretreatment methods TOC at intake (mg/L) Frequency of membrane cleaning Location
Site A (North Obhor) 1.1 Well Dual media filter with cartridge filter 0.3–0.4 Less than once every 2 years Red Sea
Site B (Corniche) 1 Well Double cartridge filter 25 and 5 μm 0.5–0.6 Every 6 months Red Sea
Site C (South Jeddah Corniche) 0.9 Well 100 μm mesh filter, antiscalant, UF and cartridge filter 0.5–0.7 Once every 6 months to one year Red Sea
Site D (Saudia plant) 0.83 Open-ocean (9 m depth) Sand filtration and cartridge filter 0.94 Every 2.5 to 3 months Red Sea
Jeddah (large capacity SWRO) 1.0 Open-ocean, nearshore Coagulation, dual media filter, chlorination/dechlorination 1.0 Originally every 2 months Red Sea
Large capacity SWRO 1.3–1.9 Open-ocean, nearshore Coagulation, dual media filter, chlorination/dechlorination 1.3–1.9 Originally monthly Arabian Gulf
Tampa SWRO 4.29–4.78 Power-plant co-shared from estuary Upward sand and precoat filtration, cartridge filter 5 μm 4.29–4.78 Every 2 months Gulf of Mexico
N.J. Harvey, et al. Desalination 496 (2020) 114735
2

Tampa Bay Water SWRO facility has had exceptional rates of biofilm
formation compared to other comparable facilities in different regions.
This research was conducted to determine if the issues are related to
solely raw water quality or are design and operational issues also
contributing to the problems. Suggested remediation based on our re-
sults is also reported in this study.
2. Materials and methods
2.1. Description of Tampa Bay Water Desalination Plant
The Tampa Bay Water Seawater Desalination Plant is located at
13041 Wyandotte Road in Gibsonton, Florida on the eastern side of
Tampa Bay, adjacent to the Teco Big Bend Power Station, a coal and
natural gas burning power plant (Fig. 1). The desalination plant pro-
vides up to 94,697 m
3
/day (25 MGD) of drinking water, which is about
10% of the service area demand [29]. Seawater is taken from Hills-
borough Bay, an estuarine system, via the cooling discharge of the
Fig. 1. Map of Tampa and surrounding counties (Florida, USA) with blue lines indicating the water supply transport routes from the Tampa Bay Water Desalination
Plant.
N.J. Harvey, et al. Desalination 496 (2020) 114735
3

power plant and treated through a variety of processes to make potable
water that meets all safe drinking water standards [30]. The seawater
goes through a variety of pretreatment screens, sand filters, and car-
tridge filters before reaching the valuable RO membranes, which have
dissolved chloride rejections near 98.6% and are the leading technology
for new desalination installations.
A full process diagram for the current Tampa Bay SWRO plant is
shown in Fig. 2. The inflow line from the power plant discharge is
periodically cleaned using chlorine dioxide. Ferric chloride is used to
coagulate suspended sediments and organics with downstream floccu-
lation into the sand filtration system. The water flows to the diato-
maceous earth precoat filter after sand filtration and into the cartridge
filter system before entering the primary membrane desalination pro-
cess. The cartridge filters were changed about 8 months prior to the
beginning of the 5-month sampling period. The plant was not operating
during most of the 8 month period.
2.2. Water sampling through the treatment system
Water samples were collected at the plant once a month from
January 2019 to May 2019; the span of time during which the plant
runs each year for supplemental water supply during the dry season for
the Tampa area. Samples were collected in 60.6 L (16-gallon) con-
tainers as well as 1-liter amber glass bottles with no headspace and
placed in ice to minimize biological activity during transport back to
the laboratory the same day. Samples were stored at 4 °C in the la-
boratory before analysis and were quickly analyzed.
The sampling sites (Fig. 2) at the plant include one of the raw water
before pretreatment screening as a reference (S1). The second sample
(S2) was taken after the upflow sand filters and the third (S3) was taken
after the water has passed through all pretreatment filters. The fourth
sample (S4) was taken after the cartridge filters and before the RO
process and the fifth sample (S5) was taken after the second pass
through the RO membranes and was treated process water (finished
water). Lastly, the sixth sample (S6) was from the concentrate stream.
2.3. Water quality measurements
2.3.1. Water quality parameters
Water quality parameters including dissolved oxygen (DO), tur-
bidity, pH, salinity, conductivity, and water temperature were mea-
sured using a ProDSS YSI probe for each sample site in the field in
triplicate.
2.3.2. Microorganism quantification
Flow cytometry has become a popular method to determine mi-
crobial cell numbers in lieu of conventional fluorescence microscopy
because it can quickly analyze the number, size, viability, and the
physiology of cells with a combination of various fluorescent dyes [31].
SYBR Green nucleic acid stain is the predominant staining solution for
use with the flow cytometer to analyze bacteria cell concentrations in
water samples and has been done with samples from Saudi Arabia and
other global locations [13,32]. Similar methods have been used to
quantify bacterial counts in samples from Lake Zurich, Switzerland [33]
and the Chungcheong province in Korea [34].
Water samples used for microbial cell counts were put into 50 mL
centrifuge tubes with 2% volume of 30% formaldehyde solution and
placed at −80 °C until analysis. Bacterial and algal cells were measured
using an Accuri C6 plus flow cytometer. Lasers were used to excite
unstained autofluorescent cells of phototrophs (mainly picocyano-
bacteria) and stained bacterial cells. Laser wavelengths were set at
488 nm for blue, green emission collected in the FL1 channel
(533 ± 30 nm) and red fluorescence in the FL3 channel (> 670 nm)
[35,36]. The flow cytometer was calibrated using 2 drops of BD™ CS&T
RUO Beads (beads consist of equal quantities of 3-μm bright, 3-μm mid,
and 2-μm dim polystyrene beads in phosphate buffered saline (PBS)
Fig. 2. Process diagram of the Tampa Desalination Facility with location numbers indicating water sampling/collection sites.
N.J. Harvey, et al. Desalination 496 (2020) 114735
4

with bovine serum albumin (BSA) and 0.1% sodium azide) in 500 μL of
ultrapure water. The frozen water samples were thawed in a beaker of
warm water for approximately 10 min before performing the analyses.
For the bacterial counts, 50 mL of each water sample was pipetted into
10 mL tubes and placed in 35 °C water to incubate for 10 min. The
samples were then stained with 5 μL of SYBR Green II RNA gel staining
solution, vortexed, and placed back into the 35 °C water to incubate for
another 10 min [35,36]. After the second incubation, each sample was
vortexed and measured on the flow cytometer individually. The system
settings used were as follows: the run limit was set to 50 μL, fluidics on
medium (35 μL/min, core size 16 um), and the threshold was set to 600
in the FL1 channel for the total bacterial cell counts [36]. For unstained
autofluorescent counting of autotrophs, 500 μL of from each sample
was pipetted into a 10 mL tube and incubated for 10 min at 35
o
C and
then processed with a run limit set to 50 μL, fluidics was set on
‘medium’ (35uL/min, core size 16 μm) and a threshold of 900 for red
fluorescence in the FL3 channel was used [36]. Each time three vials
were analyzed as triplicates.
2.3.3. Total organic carbon and organic fraction concentrations
TOC measurement involves the oxidation of carbon and the detec-
tion of carbon dioxide. It is another used method to quantify organics
passing through the treatment processes of a desalination facility. Since
pretreatment for RO membranes excludes most organic particulates,
when measuring TOC within the stages of a desalination plant, TOC is
close to that of the dissolved organic carbon (DOC) [36]. TOC mea-
surements are reported to be more sensitive than particle counting, so it
has been used in coordination with other methods such as flow cyto-
metry and UV absorbance [13,32,37,38]. Measurements made using a
Liquid Chromatography Organic Carbon Detector (LC-OCD) give fur-
ther details on the dissolved organic carbon fractions of natural organic
matter (NOM) in seawater [13,32]. Molecular masses of humics were
also studied in these reports as well as Domany et al. [38]. In order to
measure organic matter as well as keep the integrity of the study by
providing TOC measurements in addition to particle counts, the sam-
ples from the Tampa Bay Water seawater desalination plant were also
run for TOC and LC-OCD.
TOC was measured using a Shimadzu TOC-L analyzer. A volume of
20 mL from each sample was pipetted into individual tubes to allow
room for CO
2
gas bubbles to be released. A 2.1% NaCl saltwater stan-
dard solution as well as three dilutions using a carbon standard were
made (10 mg/L, 50 mg/L, and 100 mg/L) ahead of time using volu-
metric flasks and stored at 4 °C until use. Sulfuric acid (9 N H
2
SO
4
) on a
9-minute sparge time was used instead of hydrochloric acid (HCl),
which is typically used for freshwater samples, as a stronger acid is
needed for sparging seawater samples. A calibration curve was created
using the mean area (mg/L) of the standards.
The main fractions and composition of the DOC can be assessed
using a LC-OCD. The Protocols and methods by Huber et al. [39] were
followed to measure the organic fractions using a LC-OCD (LC-OCD-
OND Model 8, DOC-Labor, Germany). The samples for the LC-OCD were
prefiltered using a 0.45 μm syringe filter to exclude non-dissolved or-
ganics and a system cleaning was done by injecting 4 mL of 0.1 mol/L
NaOH through the column for 260 min. Following the cleaning steps,
2 mL of the sample was injected for analysis with 180 min of retention
time and a flow rate of 1.5 mL/min. A mobile phase of phosphate buffer
with STD 28 mmol and a pH of 6.58 was used to carry the sample
through the system. The analysis result is a chromatogram showing a
plot of signal response of different organic fractions to retention time.
Manual integration of the data was then performed to determine the
concentration of the different organic fractions based on five size
classifications that include biopolymers, humic substances, building
blocks, low molecular weight acids and low molecular weight neutrals
[13]. A summary of the composition of the LC-OCD classifications by
Villacorte [40], is given in Table 2. Perhaps the most important clas-
sifications are the biopolymers and humic substances, which cause
membrane preconditioning, leading to biofilm formation [41–43].
Some of the low molecular weight substances may also be important
[11].
Water samples for LC-OCD analysis were shipped to the King
Abdullah University of Science and Technology (KAUST) in Saudi
Arabia within three days after collection. They were shipped on ice in a
cooler and received in Saudi Arabia for further processing within 48 h
after the package was received. LC-OCD samples were placed in 40 mL
amber glass bottles and secured tightly with parafilm, bubble wrap, and
Ziplock bags to avoid contamination and spillage.
2.3.4. TEP measurements
TEP is formed by the abiotic assembly of dissolved acidic poly-
saccharides and other organic material and/or from the extracellular
excretion by algae and bacteria [45]. TEP or components of TEP in the
colloidal size range (e.g., lectin-like humic substances) commonly at-
tach to the surface of the membrane, which causes preconditioning with
the ultimate attachment of bacteria and subsequent formation of bio-
film [11,46]. Particulate TEP is defined as having a size greater than
0.4 μm while colloidal TEP has a size range from 0.1–0.4 μm [11]. Both
particulate and colloidal TEP were measured based on the method
developed by Passow and Alldredge [47]. Villacorte et al. [48] have
suggested improvements to the original method for TEP measurement
described by Passow and Alldredge [47] and Villacorte et al. [42]. The
procedure consists of filtering water through various size membranes,
staining with Alcian Blue dye, soaking in sulfuric acid, and measured on
a UV spectrometer to gain the concentration through an initial cali-
bration using Xanthan Gum, a known polysaccharide [13]. Since TEP
cannot be measured directly, it is analyzed indirectly using a calibration
curve as explained in detail by Passow and Alldredge [47] and Villa-
corte [42]. The use of this method in analysis of TEP at SWRO facilities
is explained by Dehwah and Missimer [13].
First, an Alcian Blue dye staining solution was prepared using 0.02%
Alcian Blue 8GX in a 0.06% acetate acid buffer solution with a pH ~2.5
[35,42]. The solution was then filtered through a 0.2 μm pore size
polycarbonate membrane using a vacuum pump on a constant vacuum
before use. A second solution, XG Standard, was prepared by mixing
20 mg of Xanthan Gum in 200 mL of DI water [42].
To create the calibration curve, a portion of XG Standard Solution
was diluted by 10 to ensure the consistency of the solution was thin
enough for TOC measurement and device measurement range (10 mL
XG Standard Solution + 90 mL DI). The new diluted solution was se-
quentially filtered through 0.4 μm and 0.1 μm polycarbonate mem-
branes and processed in a TOC analyzer (Shimadzu). The TOC readings
were then converted from mg/L of TOC into mg of Xanthan Gum by
dividing by the elemental carbon in the chemical formula for Xanthan
Gum (C
35
H
49
O
29
).
Separately, using 0.4 μm and 0.01 μm polycarbonate filters, dif-
ferent volumes (0 mL, 0.5 mL, 1 mL, 2 mL, and 3 mL) of the original
Xanthan Gum Standard Solution were filtered through and stained with
1 mL of Alcian Blue Dye, washed with 10 mL of DI water on constant
vacuum using a volumetric flask with filtration pieces attached. After
the initial 10 mL of DI water was filtered through, the vacuum pump
was turned off and the dye was let to soak in for approximately 10 s
before turning back on and washing the filter with the final 10 mL of DI
water to remove any remaining dye residue. The filter was then re-
moved and placed in a 50 mL glass beaker and soaked in 6 mL of 80%
sulfuric acid for 2 h, stirring occasionally. After 2 h, the acid from each
beaker was pipetted into a cuvette and placed in the UV spectrometer
(Spectronic 200) at 752 nm and the absorbance was measured. The
absorbance readings were then multiplied by the volume of Xanthan
Gum added (0 mL, 0.5 mL, 1 mL, 2 mL, and 3 mL) to calculate the mass
of Xanthan Gum on the filters. The absorbance was plotted against the
Xanthan Gum mass on each filter to find the calibration curve and the
TEP concentration was calculated.
For each sample water, 250 mL of water was passed through a
N.J. Harvey, et al. Desalination 496 (2020) 114735
5

0.4 μm filter. A volume of 200 mL of the filtered water was put in a
beaker and placed aside. The 0.4 μm filter was then washed and stained
using the same procedure as described and placed in a 50 mL beaker.
The 200 mL of product water was then used for the 0.1 μm filter fol-
lowing the same procedure for every sample and at least one duplicate
of each. It should be noted that the filtration devices were continuously
cleaned after each filter staining to avoid contamination. A volume of
6 mL of the 80% sulfuric acid was then distributed to each beaker with
filters and covered with parafilm, stirred, and let soak for 2 h and then
measured at 752 nm on the UV spectrometer to determine the absor-
bances. TEP was then calculated using the calibration curve equation.
The measurements of p-TEP and c-TEP are performed with an in-
direct measurement based on calibration to the Xanthan Gum, and are
reported as concentration equivalents of Xanthan Gum in μg Xeq./L.
These measurements must be considered to be semi-quantitative. The
duplicate samples show that the replication is good with absorbance
measurement differences occurring between 0.002 nm and 0.08 nm for
p-TEP and 0.0007 nm and 0.28 nm for c-TEP (Fig. 3).
2.3.5. Nutrients
For initial testing, 50 mL of each sample was filtered using a syringe
with an attached GF/F filter for nitrate and phosphate testing.
Additionally, 20 mL of each sample was added with 2 mL of the oxidizer
solution and then digested using persulfate digestion for simultaneous
oxidation of nitrogen and phosphorus to measure concentrations of
total nitrogen and total phosphorous through a continuous flow ana-
lyzer (Seal Model AA-3) [49]. Sample 6 was diluted by 50% for pro-
cessing due to the high salinity of the original sample.
3. Results
3.1. Fundamental water quality parameters
The water temperature, DO, conductivity, salinity, pH, and turbidity
were measured at the six sites (Table 3). The data indicate that in-
coming seawater temperature (S1) was highest at approximately 35 °C
in May while pH for the incoming seawater was highest at 8.07 in
February. The salinity generally increased during the period of sam-
pling. Turbidity remained consistently low for all filtered samples. DO
was higher for S2-S5, which all are stationed within the plant. Con-
ductivities were highest in May while January and March have very
similar conductivity values. Since the desalination facility is adjacent to
the Teco Big Bend Power Station, which withdraws and discharges up
to 5.3 million m
3
/d (1.4 billion gpd) of seawater as cooling water for
the power plant, the Tampa Bay Seawater Desalination plant diverts up
to 166,667 m
3
/d (44 MGD) of that warm seawater into the SWRO plant
for treatment. Temperature of the feedwater is affected by the power
plant cooling function and the changes in operation at the power plant.
It should be noted that the turbidity increased from the precoat filters
(S3) through the cartridge filters on several occasions. The largest in-
creases occurred in January from 0 to 5.46 NTU and in April from 0.05
to 3.41 NTU.
3.2. Microorganism quantifications
3.2.1. Algae concentrations
High algae concentrations occurred in January with declines
through the dry season from February to April (S1) (Fig. 4). In May, the
algal concentrations were the highest in the entire sampling period.
Note that algae were still present in low concentrations after the first
filtration phase during the entire sampling period (S2). No algae were
present after the second filtration (S3). The small signals found in some
of the samples in S4 to S6 were likely non-biological fluorescent par-
ticles, despite the issue that none of these particles were found in the S3
sample. The concentrations were very low and not considered to be
problematic.
3.2.2. Bacteria concentrations
Bacterial concentrations in the raw feedwater varied considerably
with the highest values occurring in April and January (Fig. 5). Note
that these values are quite high and in all cases over 1 million cells/mL.
A diagram showing the effectiveness of the pretreatment stages to re-
move bacteria is shown in Fig. 6. The upward filtration (2) did not
reduce the concentration in January and did not function to remove
bacteria. In subsequent months the upward filtration more effectively
reduced cell numbers. The second filter (S3) further reduced the cells by
6 to 72% from the upward filtration (S2). The cartridge filters had some
impact on the reduction in bacteria except for January, where it was
close to the same as in the inflow water (S4). Note that in all cases,
bacteria occur in the permeate with the highest values in January (S5).
The concentrate contained high numbers of bacteria consistent with the
bacteria concentrations passing through the cartridge filters. Since the
sampling portal for the permeate was located after chlorination, the
particles in the permeate could be dead cells of marine
Table 2
Characteristics of organic compound classifications measured by the LC-OCD techniques (modified from Villacorte [40] as defined by Huber et al. [39]
and Batsch et al. [44]).
Organic matter fraction Size range (Da) Typical composition
Biopolymers > 20,000 Polysaccharides, proteins, amino sugars, polypeptides, TEP, EPS
Humic substances ~1000 Humic and fulvic acids
Building blocks 300–450 Weathering and oxidation products of humics
LMW neutrals < 350 Mono-oligosaccharides, alcohols, aldehydes, ketones, amino acids
LMW acids < 350 All monoprotic acids
Fig. 3. Calibration curves for xanthan gum used to calculate p-TEP and c-TEP.
N.J. Harvey, et al. Desalination 496 (2020) 114735
6

ultramicrobacteria bacteria, which are known to be chlorine resistant.
These bacteria are small enough to pass through the membranes. Some
bacterial regrowth could be occurring, but it is doubtful based on the
change from a seawater to freshwater environment. The chlorine con-
centration is, however, greater than 5 mg/L. There was no sampling
portal located before chlorination.
3.3. Total organic carbon (TOC) and organic fraction concentrations
3.3.1. TOC
During the first round of sampling in January, samples were ad-
ditionally filtered through 0.2 μm filters to determine if TOC and DOC
were different. It was concluded that TOC was equal to that of DOC for
all samples as all of the organic carbon in the samples was mainly
dissolved in the water with a 0.4 mg/L average difference between DOC
and TOC samples [37].
TOC measurements remained fairly stable between 4 and 5 mg/L for
Table 3
Water quality parameters taken on site at the plant from January 2019–May 2019. Sampling location information: S1 raw water, S2 after sand upflow filters, S3 after
all primary filtration, S4 after cartridge filters, S5 finished water, S6 concentrate.
Month Site Temp (°C) DO (%) DO (mg/L) Conductivity (μS/cm) Salinity (ppt) pH Turbidity (NTU)
January S1 26 95 6.80 35,001 21 7.70 1.64
S2 26 104 7.43 35,687 22 7.81 0.20
S3 26 105 7.47 35,715 22 7.88 0.00
S4 27 105 7.37 36,255 22 7.86 5.46
S5 27 97 7.69 410.77 0 8.38 0.22
S6 28 113 6.53 82,104 54 7.64 1.97
February S1 29 94 6.36 39,280 23 8.07 3.69
S2 28 104 7.10 38,763 23 7.98 0.16
S3 28 103 7.06 38,775 23 7.96 0.00
S4 29 104 6.99 39,558 23 7.97 0.15
S5 30 97 7.31 491.43 0 8.58 0.15
S6 30 109 6.14 84,175 53 7.75 0.00
March S1 23 91 6.77 35,041 23 7.83 2.11
S2 23 100 7.56 34,958 23 7.74 0.17
S3 23 100 7.55 34,907 23 7.71 0.00
S4 23 101 7.54 35,201 23 7.70 0.00
S5 24 94 7.96 375.60 0 7.73 0.13
S6 25 100 6.00 81,837 58 7.46 0.29
April S1 30 88 5.77 42,684 25 7.90 0.76
S2 30 101 6.68 42,405 24 7.75 0.25
S3 31 97 6.31 175,038 24 7.72 0.05
S4 31 101 6.58 42,904 24 7.70 3.41
S5 32 94 6.88 431.03 0 8.36 0.00
S6 32 106 5.57 97,669 61 7.41 0.00
May S1 35 87 5.29 47,298 25 7.94 2.46
S2 34 101 6.29 46,521 25 7.88 0.28
S3 34 101 6.23 46,978 25 7.74 0.03
S4 34 101 6.24 46,570 25 7.70 0.16
S5 35 94 6.48 493.43 0 8.66 0.01
S6 36 132 5.11 101,898 59 8.05 1.87
Fig. 4. Concentrations of algae (pico- and nanophytoplankton) for each sampling site from January to May 2019. Sampling location information: S1 raw water, S2
after sand upflow filters, S3 after all primary filtration, S4 after cartridge filters, S5 finished water, S6 concentrate.
N.J. Harvey, et al. Desalination 496 (2020) 114735
7

all five sampling months, with April and May having the lowest con-
centrations and January having the highest concentration for samples
S1 to S4 (Fig. 7). With the removal of the algae and bacteria during
filtration stages at the plant, dissolved organic carbon was considered
as the primary TOC source. There is little removal of TOC from the raw
feedwater via the two filtration processes S2 and S3. The cartridge fil-
ters have a minor impact on TOC concentrations as well (S4). Some of
the TOC passed through the SWRO membranes and occurred in the
potable water in concentrations ranging from 0.13 to 0.54 mg C/L (S5)
with an average percent difference of 0.58%. As expected, the TOC
concentrations were high in the concentrate (S6). The S6 data values
may be skewed as these samples have very high salinities that could
affect the quantification process of measuring TOC as it is possible that
the acid used to sparge the samples was not strong enough for a heavily
concentrated saltwater sample.
3.3.2. Organic fractions of the TOC (LC-OCD data)
The largest percentage of organic matter occurs as humic substances
and low molecular weight neutrals (Table 4). Large fractions of bio-
polymers were found in S5 for January as well as S2 for February.
Biopolymer fractions for March, April, and May showed more con-
sistent results, while May had the lowest biopolymer fraction percen-
tages. May also had the lowest humic substance percentages at less than
20% of the water sample composition. Overall, sites S1-S6 have similar
compositions except for S5, which has a greater amount of low mole-
cular weight neutrals and low molecular weight acids with no humic
substances, indicating that the processed water has no humic sub-
stances in it as it had been removed by the RO membranes.
A more detailed assessment of the biopolymer fraction is essential
because this fraction contains the p-TEP components and other sub-
stances that are large in diameter and potentially impact the rate of
biofouling. A comparison of the biopolymer fraction in the raw water
and through each process shows that the average change in con-
centration from the raw water after the upward filtration process pro-
duces an average reduction of 17%. The precoat filters remove an
average of 14% of the concentration in the inflow. The seawater passing
Fig. 5. Concentrations of bacteria for each sampling site from January to May 2019. Sampling location information: S1 raw water, S2 after sand upflow filters, S3
after all primary filtration, S4 after cartridge filters, S5 finished water, S6 concentrate.
Fig. 6. Effectiveness of bacteria removal by the pretreatment process.
N.J. Harvey, et al. Desalination 496 (2020) 114735
8

through the cartridge filters shows an average increase of 24.5% com-
pared to the inflow. There is a considerable change in the concentration
from month to month through the filter systems preceding the SWRO
process. Note that some biopolymers occur in the permeate in all
5 months of sampling. A considerable amount of the TOC passes
through the SWRO membranes into the permeate, including some of the
building blocks, high concentrations of low molecular weight neutrals,
and some low molecular weight acids.
3.4. TEP concentrations
Both p-TEP and c-TEP were found in high concentrations in the raw
feedwater in all five months tested (Fig. 8). The concentration of p-TEP
ranged from 200 to 556 μg Xeq./L and c-TEP from 352 to 631 μg Xeq./L
in the feedwater (S1). The c-TEP had an average concentration 32%
higher than the p-TEP in the feedwater.
The upflow filters, the Precoat filter, and the cartridge filters re-
duced the concentration of p-TEP on average by 26%, 25%, and 18%,
respectively, based on a comparison of the inflow concentration to the
outflow concentrations, respectively. In March, the concentration of p-
TEP actually increased across the cartridge filter. It is important to
observe that an average p-TEP concentration of 153 μg Xeq./L passed
through the cartridge filters into the primary membrane process.
Significant concentrations of p-TEP were measured in the product water
and high concentrations in the concentrate (S6).
The c-TEP changes through the filters were inconsistent, but the
overall reductions through the upflow filters, the Precoat filter, and the
cartridge filters were 40, 0, and 4%, respectively. The average con-
centration of c-TEP that passed through the cartridge filters into the
primary membrane process was 283 μg Xeq./L. The product water
Fig. 7. Total organic carbon measurements for each site at the desalination plant. Sampling location information: S1 raw water, S2 after sand upflow filters, S3 after
all primary filtration, S4 after cartridge filters, S5 finished water, S6 concentrate.
Table 4
Organic fraction concentrations for biopolymers, building blocks, low molecular weight acids, humic substances, and low molecular weight neutrals in ppb for each
month of sampling. n.q. = not quantifiable (< 1 ppb; signal-to-noise ratio).
Sample site Sampling month Biopolymers Humic substances Building blocks Low molecular weight neutrals Low molecular weight acids
μg/L-C μg/L-C μg/L-C μg/L-C μg/L-C
Intake seawater (S1) January 89 3305 1054 2547 n.q.
February 83 2791 433 4690 248
March 175 2935 523 1496 33
April 211 2835 429 1604 6
May 35 879 n.q. 5861 119
Post upward sand filter (S2) January 64 3222 934 3820 n.q.
February 102.5
a
2818 414 4689 293
March 136 2607 458 1555 24
April 188 2547 366 1675 15
May 18.6
a
686 152 6040 19
Post precoat filter (S3) January 82 3056 640 4259 n.q.
February 65 2232 525 4641 289
March 150 2613 432 1542 23
April 186 2420 450 1737 7
May 5 591 110 6129 22
Post cartridge filter (S4) January 133 2627 1162 4276 50
February 68 2173 556 4696 267
March 143 2626 461 1484 12
April 187 2440 446 1770 n.q.
May 8 586 132 5779 6
Chlorinated product water (S5) January 262 n.q. 122 2437 15
February 3 n.q. 63 2865 6
March 4 n.q. 33 316 17
April 20 n.q. 57 421 19
May 1 n.q. n.q. 2709 6
Concentrate/brine waste water (S6) January 195 8508 1393 12,108 n.q.
February 141 5747 918 10,908 754
March 276 7011 986 3497 29
April 450 6623 817 4223 42
May 288 1511 296 13,389 70
a
Decimal error occurred in original data.
N.J. Harvey, et al. Desalination 496 (2020) 114735
9

contained significant concentrations of c-TEP and the concentrate
contained slightly elevated c-TEP concentrations.
3.5. Nutrients
Moderate concentrations of nutrients were measured in the raw
feedwater with the nitrogen analytes exceeding the phosphorus ana-
lytes (Fig. 9). The total nitrogen concentration showed a minor reduc-
tion in concentration passing through the three filtration processes with
some passing through the membranes into the permeate. The combined
nitrate and nitrite showed a similar pattern to total nitrogen with low
concentrations passing through the membrane process into the
permeate in 2 of the 5 months sampled. Total phosphorus was less than
100 μg-P/L in the raw feedwater and the phosphate as P was lower at
below 40 μg-P/L. Substantial percentages of total phosphorus and
phosphate were removed across the upflow filters. The ferric chloride
feed before the first filtration system likely impacted the reductions in
concentrations. Nearly insignificant changes in concentration of total
phosphorus and phosphate changed across the precoat and the car-
tridge filters. Some phosphorus passed through the membrane process
into the finished water.
4. Discussion
4.1. Raw feedwater quality
The feedwater for the Tampa Bay Water SWRO is exceptionally
difficult to treat. The water has a very high concentration of TOC
compared to most operating SWRO plants in the world (Table 1). Some
of the worst documented membrane biofouling issues have been
documented in the Arabian Gulf in the United Arab Emirates, Qatar,
and Saudi Arabia, where the TOC concentrations range from 0.5 to
3.9 mg/L [50,51]. Additional data on TOC in the Arabian Gulf in
coastal Saudi Arabia documented a range of 1.5 to 2.4 mg/L [51–53].
Most of the large capacity SWRO facilities on the Arabian Gulf have had
continuing issues with membrane biofouling. In the Red Sea, TOC va-
lues ranging from 0.83 to 2.95 have been reported [35,54,55]. How-
ever, membrane biofouling at lower TOC values has also been docu-
mented in the Red Sea [13,35]. In comparison, the Tampa Bay Water
feedwater has a TOC concentration consistently over 4 mg/L (Fig. 7).
In the Red Sea, there is a good database on bacteria concentrations,
which tends to range between 1.1 × 10
5
to 2.2 × 10
6
cells/mL with an
average of 5.3 × 10
5
cells/mL at SWRO intakes [35,55]. Saeed et al.
[53] reported Red Sea bacteria concentrations of 1.2 × 10
4
to
2.7 × 10
5
cells/mL and 1.3 × 10
4
to 3.3 × 10
5
cells/mL in the Arabian
Gulf. Rachman et al. [56] reported bacteria concentrations ranged
between 7.0 × 10
5
and 1.0 × 10
6
cells/mL in the Arabian Sea. They
also reported values of 7.0 × 10
5
cells/mL in the Caribbean Sea at
Providenciales, Turks and Caicos Islands, and 2.9 × 10
5
cells/mL in the
Mediterranean Sea at Alicante, Spain. In comparison, the bacteria
concentrations at the Tampa Bay Water intake peaked at about
3.5 × 10
6
cells/mL and the average was about 1.8 × 10
6
cells/mL,
which was quite high in comparison to the other locations.
TEP and the biopolymer fractions of organic matter are also major
factors in the cause of the rapid biofouling of membranes [51,57]. In
the Red Sea, p-TEP ranges in concentration from 53 to 347 μg Xeq./L
with an average of 191 μg Xeq./L, and c-TEP ranges in concentration
from 36 to 287 μg Xeq./L and averages 125 μg Xeq./L [54]. In com-
parison, Tampa Bay Water feed contains p-TEP at a concentration range
of 200 to 556 μg Xeq./L with an average of 336 μg Xeq./L. The bio-
polymer concentration at the Red Sea seawater intakes ranges from 28
to 164 μg/L and averages 62 μg/L. At Tampa Bay, the range in bio-
polymer concentrations ranges from 35 to 211 μg/L with an average of
119 μg/L.
The occurrence of phosphate in the raw seawater has been sug-
gested to have an impact on the rate of membrane biofouling [16,17].
Phosphate in the raw feedwater ranged between 11.8 and 39.8 μg-P/L
(Fig. 9). The range in phosphate data from the Arabian Gulf was 1.72 to
5.20 μg-P/L and from the Red Sea was 0.03 to 0.23 μg-P/L [54]. The
high values found in the Arabian Gulf are likely influenced by domestic
wastewater discharge and may not be comparable, especially with
consideration that the reported nitrate values are 1.49 to 4.99 μg-N/L
and ammonia concentrations range from 0.79 to 8.30 μg-N/L [58].
Biological productivity is also impacted by water temperature,
which may in turn, impacts the growth rate of a membrane biofilm
[18]. Temperature data from 2001 taken by the Environmental Pro-
tection Commission of Hillsborough County shows average sea surface
temperature to be 23.5 °C while in 2010 the average was 25.5 °C [59].
The data for May in this study shows higher temperatures at 34.7 °C on
average for surface water, respectively, as cooling water from the TECO
power plant is utilized, while the temperature taken in March was the
only sample month within the range of measurements taken from 2001
to 2010. The water temperature entering the system during the sam-
pling period ranged from 22.9 to 34.7 °C and averaged 28.7 °C. The
water temperature in the permeate ranges from 23.8 to 35.2 °C and
averages 29.6 °C. The temperature is higher through the primary pro-
cess where the biofilm occurs compared to the inflow temperature.
Based on all of the physical, chemical, and biological parameters
assessed, the feedwater quality into the Tampa Bay Water SWRO plant
is perhaps the worst or one of the worst in the world in terms of
treatment difficulty. While the rate of biofouling is not as great as some
other systems documented in the MiddleEast, it is still very high.
Fig. 8. Monthly changes in colloidal and particulate TEP concentrations. Sampling location information: S1 raw water, S2 after sand upflow filters, S3 after all
primary filtration, S4 after cartridge filters, S5 product water, S6 concentrate.
N.J. Harvey, et al. Desalination 496 (2020) 114735
10

Nevertheless, there is a need to implement a pretreatment system at
Tampa Bay that can effectively remove organic substances commonly
associated with biofouling.
4.2. Effectiveness of pretreatment system
The pretreatment systems used at the plant, the old one and the
replacement system (remedial), must be classified as conventional ap-
proaches to pretreatment involving coagulation and filtration before
the primary SWRO process [14,60]. In most regions of the world with
“typical” seawater having moderate to low TOC values and limited
concentrations of algae, bacteria, and extracellular organic substances,
this approach would be effective and yield sufficient improvement in
water quality as to limit the rate of membrane biofouling and required
cleaning frequency [61]. Based on the data collected, the pretreatment
at the Tampa Bay Water SWRO plant is not very effective.
To support this conclusion, the patterns of concentration changes
from the raw feedwater across each process show a consistent pattern.
The algae show a large reduction in concentration based on the coa-
gulation with ferric chloride and upflow sand filtration, but persist in
small concentrations to and through the cartridge filters (Fig. 4). Bac-
teria are removed to a large degree by the two primary filtration sys-
tems, but still pass through the cartridge filter at concentrations ranging
from 4.0 × 10
4
to 1.1 × 10
6
cells/mL (Fig. 5).
Perhaps the most important issue concerns TEP passage through the
pretreatment system. Particulate TEP is known to influence membrane
biofouling because it preconditions the surface to allow bacterial at-
tachment and growth [41,43,57,62]. In a recent paper, Winters et al.
[11] suggested that p-TEP may be less important in the biofouling
process because it is mostly removed during pretreatment and there-
fore, some other lower molecular weight substances, such as lectin-like
humic substances may be more important. However, at SWRO facilities
where p-TEP is not being effectively removed, it is a significant factor in
the biofouling process.
In the case of the Tampa Bay Water SWRO plant, the p-TEP is not
being effectively removed and is likely a major factor in the occurrence
of membrane biofouling. An average concentration of 153 μg Xeq./L
passed through the cartridge filters into the primary membrane process
and p-TEP has even been found in the permeate (Fig. 8). The p-TEP data
are confirmed by the concentrations of the biopolymer fraction of the
organic matter, which show a passage of 8 to 187 μg/L through the
cartridge filters into the primary process. It is not known whether c-TEP
also has some influence on biofouling. Furthermore, low molecular
weight organics (especially LMW neutrals), referred to as assimilable
organic carbon, can pass through most of the pretreatment processes
and contribute to biofouling. These low molecular weight organics will
Fig. 9. Monthly changes in nutrient concentrations after step of the water treatment. The concentrations are shown for total nitrogen as N, NOx (NO
2
+ NO
3
) as N.
The phosphate (orthophosphate) concentration is lower compared to the total phosphorus. Both are expressed as phosphorus. The TN and TP for January S1 sample
are unavailable as well as TP for May S5 sample. Sampling location information: S1 raw water, S2 after sand upflow filters, S3 after all primary filtration, S4 after
cartridge filters, S5 permeate, S6 concentrate.
N.J. Harvey, et al. Desalination 496 (2020) 114735
11

even pass through low-permeability nanofiltration membranes [63].
LC-OCD analysis showed that a large amount of LMW neutrals were
able to pass through the pretreatment process and reach the RO
membrane (Table 4), which would eventually contribute to the growth
of bacteria on the RO membrane.
Another observed issue is the increase in turbidity across the car-
tridge filters observed in 4 of the 5 months with two large increases in
January (0 to 5.46 NTU) and April (0.03 to 3.41 NTU) (Table 3). This
corresponds to high bacteria passage through the cartridge filters in
January (> I0
6
cells/mL) and April (4 × 10
5
cells/mL) (Fig. 5). The
biopolymer fraction of NOM also increased in January from 82 to
133 μg/L and in April showed no change from 186 to 187 μg/L. The
concentration of c-TEP increased across the cartridge filters in January
from 200 to 390 μg Xeq./L and showed little change in April going from
220 to 200 μg Xeq./L. The p-TEP also showed high passage rates in
January and April at about 140 and 160 μg Xeq./L. These data suggest
that there may be bacterial regrowth in the cartridge filters and some
turbidity passage, perhaps as aggregated particles sloughing off. The
increase or maintenance of the biopolymer fraction suggests biological
activity. In addition to these issues, it was observed that the dose of
ferric chloride in the system does periodically bypass the two filtration
processes and the color can be observed entering the cartridge filters.
This issue cannot be verified to influence the turbidity increase.
The pretreatment processes are also ineffective in the removal of
TOC and phosphate. The TOC passes into the membrane process at over
3.5 mg/L and the phosphate at 1.2 to 9.0 μg/L. Some substances within
the TOC (e.g., lectin-like substances) and the phosphate likely also have
some effect on the biofouling rate [64].
We observed that the total nitrogen in April increased from the raw
water across the two initial filtration processes and then declined
slightly across the cartridge filters (Fig. 9). However, the concentration
of NOx showed a decreasing trend in April across the initial filtration
steps. These results suggest the observed increase in total nitrogen is
most likely due to increased concentration of ammonia. Although the
exact reason for the observed increase in total nitrogen is unknown, it is
possible that the seawater characteristics (temperature, DO, and con-
ductivity) interfere with the microbial oxidation of ammonia [65].
4.3. Options for improving the pretreatment process
Based on the very poor quality of the feedwater, there are a few
options that could be considered to better protect the primary mem-
brane process. The conventional pretreatment design could be replaced
with a combination of dissolved air floatation (DAF) followed by ul-
trafiltration (UF).
The use of combined DAF and UF (Fig. 10) would likely provide
higher quality water into the SWRO membranes and provide a greater
degree of protection for the membrane process during harmful algal
blooms, which are occurring in this region at an increased frequency
[66,67]. Investigations using different feed water quality showed that
DAF could effectively remove organic material. DAF combined with an
optimal dose of ferric chloride and pH of 5.5 significantly reduced the
biopolymers, humic substances, building blocks and LMW acids.
However, the removal of LMW neutrals was found to be nominal [68].
The addition of ferric chloride during the DAF process will facilitate the
removal of virtually all of the algae and some of the bacteria and larger
organic molecules. DAF process with built in filtration has been suc-
cessfully implemented in Singapore in locations with frequent occur-
rence of algal blooms [69].
Because of the high concentrations of bacteria and organic mole-
cules (especially TEP) found in the feedwater, the UF process should be
added as a second process for further polishing of DAF effluent.
Although UF membranes do not effectively remove all dissolved organic
carbon, the process will filter out most of the bacteria and suspended
solids, the p-TEP, and some of the large organic molecules [70–71].
Much of the c-TEP and the low molecular weight acids and neutrals will
not be removed by the UF [72]. However, the biofouling issue will be
transferred from the primary process to the UF. But, it is proven at a
commercial scale that UF fouling is better controlled through mem-
brane backwash with less chemical use [56]. Furthermore, advance-
ments in membrane technology have led to the production of mem-
branes that are comparatively more resistant to biofouling [73]. Using
such membranes would require less cleaning and reduce the plant
downtime during the cleaning procedure. However, the question re-
mains how easily and frequently UF membranes would require cleaning
under severe operating conditions though that it remains practical to
backwash with chlorinated water [74]. This would require further in-
vestigation using conventional and novel membrane cleaning methods
[75]. Tampa Bay Water is currently planning a major expansion of the
SWRO plant to 189,394 m
3
/d (50 MGD). UF pretreatment is being
considered as part of the pretreatment train and may be pilot tested.
4.4. Alternative supply of seawater: groundwater use potential
At a depth on about 320 m below the Tampa Bay Water site, a high
permeability source of saline water occurs that could be used as a
primary water source for the SWRO plant [75–77]. High-capacity wells
could be constructed into the Avon Park Formation high transmissivity
zone that occurs directly beneath the site. These wells should have a
capacity of up to 20,000 m
3
/d based on the very high transmissivity of
the production zone. The salinity of this aquifer is known to be very
close to normal seawater at about 36 g/L [75,76].
Use of the Avon Park Formation high transmissivity zone would be
feasible, but the TDS of water in this aquifer is close or slightly above
typical seawater. The efficiency of conversion could be lower compared
to the use of the surface water from Tampa Bay because it has an
average TDS of about 26 g/L compared to the aquifer water at 35 to
36 g/L [75,76]. However, the aquifer water would require minimal
pretreatment, perhaps only acidification. Feedwater from wells is
known to provide a lower operating cost when comparing similar TDS
concentrations [78]. The use of seawater from this aquifer was con-
sidered in the planning stage of the facility but was not used based on
overall regulatory restrictions that were in effect on the use of any
water from the Floridan Aquifer System. This issue is not relevant in the
pumping of the seawater from the lowest part of the aquifer, because it
will not impact the freshwater part of the aquifer system.
Fig. 10. Simplified schematic of dissolved air floatation
(DAF) followed by ultrafiltration (UF) process not showing
chemical injection, e.g., coagulants. Seawater enters into
the DAF system, which removes particles and large organics
including algae and some bacteria, and some large organic
molecules. The feedwater is further treated with UF mem-
brane that removes bacteria, TEP, and some of the other
organic molecules water before the water enters the car-
tridge filters and membranes.
N.J. Harvey, et al. Desalination 496 (2020) 114735
12

5. Conclusion
The Tampa Bay Water SWRO plant has experienced operation dif-
ficulties due to low feed water quality. The pretreatment design was
changed to help provide better quality water to the primary membrane
process. However, despite the change in the pretreatment design, the
quality of seawater continues to pose a challenge by requiring mem-
brane cleanings every eight weeks of run time.
Feedwater to the plant is taken from a biologically-productive es-
tuarine system via a power plant cooling water stream. This water has
high concentrations of TOC in the form of primarily DOC, and high
concentrations of bacteria, TEP, the biopolymer fraction of NOM, and
phosphate. The feedwater has a very high organic carbon concentration
compared to virtually all other large-capacity SWRO plants in the
world. This requires extreme pretreatment compared to other facilities.
The conventional approach to pretreatment at the facility does not
protect the primary membrane process from excessive membrane bio-
fouling. It is unlikely that any changes in operating protocol would
improve the feedwater to slow the biofouling rate. A change of the
pretreatment system to DAF followed by UF could reduce the influx of
p-TEP and other substances that accelerate biofouling. However, the UF
system would likely biofoul quickly and require frequent cleaning, but
its control remains practical through membrane backwash. Perhaps a
groundwater source should be considered as an alternative feedwater
source.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
Acknowledgments
Funding for this research was provided by the Emergent Technology
Institute, U. A. Whitaker College of Engineering, Florida Gulf Coast
University in Fort Myers, Florida. The research reported in this paper
was also supported by the Water Desalination and Reuse Center at King
Abdullah University of Science and Technology (KAUST), Saudi Arabia.
The authors thank Tampa Bay Water for allowing monthly sample
collections at the desalination plant and donating their time and effort
into this project. The authors would like to thank Dr. Abdullah Dahwah
and Haruka E. Urakawa for contributing their efforts into this research
as well with their expertise in TEP methodology and nutrient analysis,
respectively.
References
[1] L.F. Greenlee, D.F. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis
desalination: water sources, technology, and today's challenges, Water Res. 43 (9)
(2009) 2317–2348.
[2] M. Elimelech, W.A. Phillip, The future of seawater desalination: energy, technology,
and the environment, Science 333 (6043) (2011) 712–717.
[3] G. Amy, N. Ghaffour, Z. Li, L. Francis, R.V. Linares, N. Chung, T.M. Missimer,
S. Lattemann, Membrane-based seawater desalination: past, present, and future,
Desalination 401 (2017) 16–21.
[4] M. Qasim, M. Badrelzaman, N.N. Darwish, N.A. Darwish, N. Hilal, Reverse osmosis
desalination: a state-of-the-art review, Desalination 459 (2019) 59–104, https://
doi.org/10.1016/j.desal.2019.02.008.
[5] H. Frank, K.E. Fussmann, E. Rahav, E. Bar Zeev, Chronic effects of brine discharge
form large-scale seawater reverse osmosis desalination facilities on benthic bac-
teria, Water Res. 151 (2019) 478–487.
[6] H.F. Ridgeway, H.C. Flemming, Membrane biofouling, in: J. Malleviatte,
P.E. Odendaal, M.R. Wiesner (Eds.), Water Treatment Processes, McGraw-Hill, New
York, 1996, pp. 6.1–6.52.
[7] H.C. Flemming, G. Schaule, T. Griebe, J. Schmitt, A. Tamachkiarowa, Biofouling –
the Achilles heel of membrane processes, Desalination 113 (1997) 215–225.
[8] J.S. Baker, L.Y. Dudley, Biofouling in membrane systems – a review, Desalination
118 (1998) 81–90.
[9] H.S. Vrouwenvelder, J.A.M. van Passen, H.C. Folmer, J.A.M.H. Hofman,
M. Nederlof, M.D. van der Kooij, Biofouling of membranes for drinking water
production, Desalination 118 (1–3) (1998) 157–166.
[10] T. Nguyen, F.A. Roddick, L. Fan, Biofouling of water treatment membranes: a re-
view of the underlying causes, monitoring techniques and control measures,
Membrane 2 (2012) (2012) 804–840.
[11] H. Winters, T.H. Chong, A.G. Fane, W. Krantz, M. Rzechowicz, N. Saeidi, The in-
volvement of lectins and lectin-like humic substances in biofilm formation – is TEP
important? Desalination 399 (2016) 61–68.
[12] S. Negati, S.A. Mirbagheri, D.M. Warsinger, M. Fazeli, Biofouling in seawater re-
verse osmosis (SWRO): impact of module geometry and mitigation with ultra-fil-
tration, J. Water Process Engin. 29 (2019) 100782, , https://doi.org/10.1016/j.
jwpe.2019.100782.
[13] A.H.A. Dehwah, T.M. Missimer, Subsurface intake systems: green choice for im-
proving feed water quality at SWRO desalination plants, Jeddah, Saudi Arabia,
Water Res. 88 (2016) 216–224, https://doi.org/10.1016/j.watres.2015.10.011.
[14] N. Voutchkov, Pretreatment for Reverse Osmosis Desalination, Elsevier,
Amsterdam, 2017.
[15] J. Kavitha, M. Rajalakshmi, A.R. Phani, M. Padaki, Pretreatment processes for
seawater reverse osmosis desalination – a review, J. Water Process Engineer. 32
(2019) 100926, , https://doi.org/10.1016/j.jwpe.2019.100926.
[16] J.D. Jacobson, M.D. Kennedy, G. Amy, J.C. Schippers, Phosphate limitation in re-
verse osmosis: an option to control biofouling? Desalin. Water Treat. 5 (2009)
198–206.
[17] J.S. Vrouwenvelder, F. Beyer, K. Dahmani, N. Hasan, G. Galjaard, J. Kruithof,
M.V. Loosdrecht, Phosphate limitation to control biofouling, Water Res. 44 (11)
(2010) 3454–3466, https://doi.org/10.1016/j.watres.2010.03.026.
[18] X. Jin, A. Jawor, S. Kim, E.M. Hoek, Effects of feed water temperature on separation
performance and organic fouling of brackish water RO membranes, Desalination
239 (1–3) (2009) 346–359, https://doi.org/10.1016/j.desal.2008.03.026.
[19] H. Winters, Identification of critical flux and cross-flow conditions for the control of
bacterial and organic fouling of seawater reverse osmosis membranes, The Middle
East Desalination Research Center, MEDR Project 97-AS-004B, 2001.
[20] J.S. Vrouwenvelder, J.A.M. van Passan, J.M.C. van Agtmaal, M.C.M. van
Loosdrecht, J.C. Kruithof, A critical flux to avoid biofouling of spiral wound na-
nofiltration and reverse osmosis membranes: fact or fiction, J. Membr. Sci. 326 (1)
(2009) 36–44.
[21] S. Adham, P. Gagliardo, D. Smith, D. Ross, K. Gramith, R. Trussell, Monitoring the
integrity of reverse osmosis membranes, Desalination 119 (1) (1998) 143–150,
https://doi.org/10.1016/S0011-9164(98)00134-9.
[22] Z. Yang, Y. Zhou, Z. Feng, X. Rui, T. Zhang, Z. Zhang, A review on reverse osmosis
and nanofiltration membranes for water purification, Polymers 11 (2019) 1252.
[23] H. Rand, Tampa Bay Desalination Project - an unfinished story, Retrieved from,
2003. https://www.waterworld.com/international/desalination/article/
16200492/tampa-bay-desalination-project-an-unfinished-story.
[24] J. Landers, Tampa Bay Water hires team to revamp troubled desalination plant, Civ.
Eng. 75 (1) (2005) 30 Retrieved from https://search.proquest.com/docview/
228458496.
[25] J. Hughes, C. Rosenfeld, Tampa Bay Water Desalination Plant, Environmental
Finance Center at the University of North Carolina, Chapel Hill, NC, 2016 Retrieved
from https://efc.sog.unc.edu/sites/default/files/2017/Tampa%20Bay%20Water_
Final_Web.pdf.
[26] S. Behn, M. Diffenderfer, Looking out to sea: resurging interest in seawater desa-
lination projects face the twin specters of technical challenges and permitting
complexity, Retrieved from, 2015. https://www.llw-law.com/articles/looking-sea-
resurging-interest-seawater-desalination-projects-face-twin-specters-technical-
challenges-permitting-complexity/.
[27] Acconia Agua, Tampa Bay Desalination Plant, Retrieved January 11, 2020, from,
2020. https://www.acciona.us/projects/water/desalination-plants/tampa-bay-
desalination-plant/.
[28] N. Simeonova, Long Time Coming, Retrieved from, 2008. https://www.wwdmag.
com/desalination/long-time-coming.
[29] Tampa Bay Water, Water Quality Report 2018: Tampa Bay Water, https://www.
tampabaywater.org/documents/quality/Water-Quality-Report-2018.pdf, (2018).
[30] Tampa Bay Water, Tampa Bay Water reliably provides clean, safe water to the re-
gion now and for future generations, https://www.tampabaywater.org/Portals/0/
documents/supplies/Water-Quality-Report-2017.pdf, (2017).
[31] F. Hammes, T. Broger, H.-U. Weilenmann, M. Vital, J. Helbing, U. Bosshart,
P. Huber, R. Odermatt, B. Sonnleitner, Development and laboratory-scale testing of
a fully automated online flow cytometer for drinking water analysis, Cytometry Part
A 81A (6) (2012) 508–516, https://doi.org/10.1002/cyto.a.22048.
[32] A.H.A. Dehwah, S. Li, S. Al-Mashharawi, H. Winters, T.M. Missimer, Changes in
feedwater organic matter concentrations based on intake type and pretreatment
processes at SWRO facilities, Red Sea, Saudi Arabia, Desalination 360 (2015)
19–27, https://doi.org/10.1016/j.desal.2015.01.008.
[33] F. Hammes, M. Berney, Y. Wang, M. Vital, O. Köster, T. Egli, Flow-cytometric total
bacterial cell counts as a descriptive microbiological parameter for drinking water
treatment processes, Water Res. 42 (1) (2008) 269–277, https://doi.org/10.1016/j.
watres.2007.07.009.
[34] J.W. Park, Y.J. Lee, A.S. Meyer, I. Douterelo, S.K. Maeng, Bacterial growth through
microfiltration membranes and NOM characteristics in an MF-RO integrated
membrane system: lab-scale and full-scale studies, Water Res. 144 (2018) 36–45,
https://doi.org/10.1016/j.watres.2018.07.027.
[35] A.H. Alshahri, A.H.A. Dehwah, T. Leiknes, T.M. Missimer, Organic carbon move-
ment through two SWRO facilities from source water to pretreatment to product
with relevance to membrane biofouling, Desalination 407 (2017) 52–60, https://
doi.org/10.1016/j.desal.2016.12.015.
[36] R. Van der Merwe, F. Hammes, S. Lattemann, G. Amy, Flow cytometric assessment
N.J. Harvey, et al. Desalination 496 (2020) 114735
13

of microbial abundance in the near-field of seawater reverse osmosis concentrate
discharge, Desalination 343 (2014) 208–216.
[37] E.R. Ostarcevic, J. Jacangelo, S.R. Gray, M.J. Cran, Current and emerging techni-
ques for high-pressure membrane integrity testing, Membranes 8 (3) (2018) 60,
https://doi.org/10.3390/membranes8030060.
[38] Z. Domany, I. Galambos, G. Vatai, E. Bekassy-Molnar, Humic substances removal
from drinking water by membrane filtration, Desalination 145 (1) (2002) 333–337,
https://doi.org/10.1016/S0011-9164(02)00432-0.
[39] S.A. Huber, A. Balz, M. Abert, W. Pronk, Characterization of aquatic humic and non-
humic matterwith size-exclusion chromatography – organic carbon detection –
organic nitrogen detection (LC-OCD-OND), Water Res. 45 (2011) 879–885.
[40] L.O. Villacorte, Liquid chromatography – organic carbon detection (LC-OCD), in:
E. Drioli, L. Giorno (Eds.), Encylopedia of Membranes, Springer-Verlag, Berlin,
2014, , https://doi.org/10.1007/978-3-642-40872-4_714-6.
[41] T. Berman, M. Holenberg, Don’t fall foul of biofilm through high TEP levels,
Filtration & Seperation (2005) 30–32 May.
[42] L.O. Villacorte, M.D. Kennedy, G.L. Amy, J.C. Schippers, Measuring transparent
exopolymer particles (TEP) as indicator of the (bio)fouling potential of RO feed
water, Desalin. Water Treat. 5 (1–3) (2009) 207–212, https://doi.org/10.5004/
dwt.2009.587.
[43] E. Bar-Zeev, U. Passow, C.S. Romero-Vergas, M. Elimelech, Transparent exopolymer
particles (TEP): from aquatic environments and engineered systems to membrane
biofouling, Environ. Sci. Technol. 49 (2) (2015) 691–707.
[44] A. Batsch, D. Tyszler, A. Brügger, S. Panglisch, M. Thomas, Foulant analysis of
modified membranes for water and wastewater treatment with LC-OCD,
Desalination 178 (2005) 63–72.
[45] S. Li, H. Winters, L.O. Villacorte, Y. Ekowan, A.H. Emwas, M.D. Kennedy, G.L. Amy,
Compositional similarities and differences between transparent exopolymer parti-
cles (TEP) from two marine bacteria and two marine algae, Mar. Chem. 174 (2015)
131–140.
[46] T. Berman, Transparent exopolymer particles as critical agents in aquatic biofilm
formation: implications for desalination and water treatment, Desal. Water Treat.
51 (4–6) (2013) 1014–1020.
[47] U. Passow, A.L. Alldredge, A dye-binding assay for the spectrophotometric mea-
surement of transparent exopolymer particles (TEP), Limnol. Oceanogr. 40 (7)
(1995) 1326–1335, https://doi.org/10.4319/lo.1995.40.7.1326.
[48] L.O. Villacorte, Y. Ekowati, H.N. Calix-Ponce, J.C. Schippers, G.L. Amy,
M.D. Kennedy, Improved method for measuring transparent exopolymer particles
(TEP) and their precursors in fresh and saline water, Water Res. 70 (2015) 300–312,
https://doi.org/10.1016/j.watres.2014.12.012.
[49] M. Hosomi, R. Sudo, Simultaneous determination of total nitrogen and total
phosphorus in freshwater samples using persulfate digestion, Int. J. Environ. Stud.
27 (3–4) (1986) 267–275, https://doi.org/10.1080/00207238608710296.
[50] H.I. Emara, Total organic carbon content in the waters of the Arabian Gulf, Environ.
Int. 24 (1–2) (1998) 97–103.
[51] S. Li, S. Sinha, T. Leiknes, G.L. Amy, N. Ghaffour, Evaluation of potential particu-
late/colloidal TEP foulants on a pilot scale SWRO desalination study, Desalination
393 (2016) 127–134.
[52] M.I.M. Ershath, M.A. Namazi, M.O. Saeed, Effect of cooling water chlorination on
entrained selected copepod species, Biocatalysis and Agr. Biotech. 17 (2019)
129–134.
[53] M.O. Saeed, M.I. Al-Otaibi, M.I.M. Ershath, Fungal and marine shell fouling in
desalination plants equipment, J. Water Reuse Desal. 9 (4) (2019) 423–430,
https://doi.org/10.2166/wrd.2019.02.026.
[54] M.O. Saeed, M.I.M. Ershath, I.A. Al-Tisan, Perspective on desalination discharges
and coastal environments of the Arabian Peninsula, Mar. Environ. Res. 145 (2019)
1–10, https://doi.org/10.1016/j.marenvres.2019.02.005.
[55] A.H.A. Dehwah, D.M. Anderson, S. Li, F.L. Mallon, Z. Batang, A.H. Alshahri,
S. Tsegaye, M. Hegy, T.M. Missimer, Understanding transparent exopolymer par-
ticle occurrence and interaction with algae, bacteria, and fractions of natural or-
ganic matter in the Red Sea: implications for seawater desalination, Desal. Water
Treatment 192 (2020) 78–96 (in press).
[56] R.M. Rachman, S. Li, T.M. Missimer, SWRO feedwater quality improvement using
subsurface intakes in Oman, Spain, Turks & Caicos, and Saudia Arabia, Desalination
151 (2014) 88–100.
[57] S. Kerdi, A. Qamar, A. Alpatova, Equation chapter 1 section 1 N. Ghaffour, an in-situ
technique for the direct structural characterization of biofouling in membrane fil-
tration, J. Membr. Sci. 583 (2019) 81–92.
[58] A.M. Mohorjy, A.M. Khan, Preliminary assessment of water quality along the Red
Sea coast near Jeddah, Saudi Arabia, Water Int. 31 (1) (2006) 109–115.
[59] D.J. Karlen, Surface Water Quality 2001–2010: Environmental Protection
Commission of Hillsborough County, http://www.epchc.org/home/
showdocument?id=552, (2014).
[60] R. Valavala, J. Sohn, J. Han, N. Her, Y. Yoon, Pretreatment in reverse osmosis
seawater desalination: a short review, Env. Engin. Res. 16 (4) (2011) 205–211.
[61] S. Li, S.-T. Lee, S. Sinha, T. Leiknes, G.L. Amy, N. Ghaffour, Transparent exopolymer
particles (TEP) removal efficiency by a combination of coagulation and ultra-
filtration to minimize membrane fouling, Water Res. 102 (2016) 485–493.
[62] S. Meng, M.E. Rzechowicz, H. Winters, A.G. Fane, L. Yu, Transparent exopolymer
particles (TEP) and their potential effect on membrane biofouling, Appl. Microbiol.
Biotechnol. 97 (13) (2013) 5705–5710.
[63] S. Meylan, F. Hammes, J. Traber, E. Salhi, U. Vongunten, W. Pronk, Permeability of
low molecular weight organics through nanofiltration membranes, Water Res. 41
(17) (2007) 3968–3976, https://doi.org/10.1016/j.watres.2007.05.031.
[64] C.-M. Kim, S.-J. Kim, L.H. Kim, M.S. Shin, H.-W. Yu, I.S. Kim, Effects of phosphate
limitation in feed water on biofouling in forward osmosis (FO) process, Desalination
349 (2014) 51–59.
[65] B. Liu, L. Gu, X. Yu, G. Yu, H. Zhang, J. Xu, Dissolved organic nitrogen (DON)
profile during backwashing cycle of drinking water biofiltration, Sci. Total Environ.
414 (2012) 508–514.
[66] Y. Sutova, B.L. Karna, A.C. Hambly, B. Lau, R.K. Henderson, P. Le-Clech, Enhancing
organic matter removal in desalination pretreatment systems by application of
dissolved air flotation, Desalination 383 (2016) 12–21.
[67] Y. Shutova, B.L. Karna, A.C. Hambly, B. Lau, R.K. Henderson, P. Le-Clech,
Enhancing organic matter removal in desalination pretreatment systems by appli-
cation of dissolved air floatation, Desalination 383 (1) (2016) 12–21.
[68] F.H. Kiang, W.W.L. Young, D.D. Ratnayaka, The Singapore solutions, Civ. Eng. 77
(1) (2007) 62–69.
[69] J. Zhang, S. Gao, H. Feng, F. Zhanh, C. Li, Y. Liu, D. Fu, C. Ye, Pilot testing of
outside-in UF pretreatment prior to RO for high turbidity seawater desalination,
Desalination 196 (1–3) (2006) 66–75.
[70] A.R. Guastalli, F.X. Simon, Y. Penru, A. de Kerchove, J. Llorens, S. Baig, Comparison
of DMF and UF pretreatments for particulate material and dissolved organic matter
removal in SWRO desalination, Desalination 322 (2013) 144–150.
[71] O. Lorain, B. Hersant, F. Persin, A. Grasmick, N. Brunard, J.M. Espenan,
Ultrafiltration membrane pretreatment benefits for reverse osmosis process in
seawater desalting: quantification in terms of capital investment cost and operating
cost reduction, Desalination 203 (1–3) (2007) 277–285, https://doi.org/10.1016/j.
desal.2006.02.022.
[72] S. Konagaya, K. Nita, Y. Matsui, M. Miyagi, New chlorine-resistant polyamide re-
verse osmosis membrane with hollow fiber configuration, Appl. Polym. Sci. 79 (3)
(2001) 517–527.
[73] A. Alpatova, A. Qamar, M. Al-Ghamdi, J.G. Lee, N. Ghaffour, Effective membrane
backwash with carbon dioxide under severe fouling and operation conditions, J.
Membr. Sci. 611 (2020) 118290.
[74] M. Al-Ghamdi, A. Al-Hadidi, N. Ghaffour, Membrane backwash cleaning using CO
2
nucleation, Water Res. 165 (2020) 114985.
[75] J.J. Hickey, Hydrogeological data for the South Cross Bayou subsurface-injection
test site, Pinellas County, Florida, U. S. Geological Survey Open-file Report 78-575,
Tampa, FL, USA, 1979.
[76] J.J. Hickey, Hydrogeology and results of injection tests at waste-injection test sites
in Pinellas County, Florida, U. S. Geological Survey Water-supply Paper 2183, 1982
(41 pp, Tampa, FL, USA).
[77] P.D. Ryder, Hydrology of the Floridan Aquifer System in west-central Florida, U. S.
Geological Survey Professional Paper 1403-F, Washington, D. C. 1985.
[78] T.M. Missimer, N. Ghaffour, A.H.A. Dehwah, R. Rachman, R.G. Maliva, G.L. Amy,
Subsurface intakes for seawater reverse osmosis systems: capacity limitation, water
quality improvement, and economics, Desalination 322 (2013) 37–51.
N.J. Harvey, et al. Desalination 496 (2020) 114735
14