Aquifer Treatment of Sea Water to Remove
Natural Organic Matter Before Desalination
by Abdullah H. A. Dehwah1 , Samir Al-Mashharawi1 , Kim Choon Ng1 , and Thomas M. Missimer2
Abstract
An investigation of a sea water reverse osmosis desalination facility located in western Saudi Arabia has shown that aquifer
treatment of the raw sea water provides a high degree of removal of natural organic matter (NOM) that causes membrane biofouling.
The aquifer is a carbonate system that has a good hydraulic connection to the sea and 14 wells are used to induce sea water
movement 400 to 450 m from the sea to the wells. During aquifer transport virtually all of the algae, over 90% of the bacteria,
over 90% of the biopolymer fraction of NOM, and high percentages of the humic substance, building blocks, and some of the
low molecular weight fractions of NOM are removed. Between 44 and over 90% of the transparent exopolymer particles (TEP)
are removed with a corresponding significant reduction in concentration of the colloidal fraction of TEP. The removal rate for TEP
appears to be greater in carbonate aquifers compared to siliciclastic systems. Although the production wells range in age from
4 months to 14 years, no significant difference in the degree of water treatment provided by the aquifer was found.
Introduction
Unconfined aquifer systems have been used for more
than a century in the water treatment process to improve
water quality in potable-water systems, such as riverbank
filtration (Ray et al. 2002; Hubbs 2005), or for at the last
40 years to remove contaminants in infiltrated domestic
waste water which is termed soil-aquifer treatment
(Drewes et al. 2003). The groundwater system is known
to produce a robust degree of water treatment caused
by several processes including filtration, absorption,
adsorption, dilution, and biochemical degradation. Use
of shallow aquifer systems to provide feed water to sea
water desalination plants is becoming another important
application of aquifer treatment (Missimer 2009; Missimer
et al. 2013).
Potable water deficits in many regions of the
world are being met by the creation of “new” fresh
water by desalting sea water. Of the various processes
used for desalination, the commercially most economic
1
Water Desalination and Reuse Center, King Abdullah
University of Science and Technology, Thuwal 23955-6900, Saudi
Arabia.
2
Corresponding author: U. A. Whitaker College of Engineering,
Florida Gulf Coast University, 10561 FGCU Blvd. South, Fort Myers,
FL 33965-6565; [email protected]
Article impact statement: Aquifer treatment has been
demonstrated to provide a high degree of pretreatment to remove
natural organic matter from sea water.
Received January 2016, accepted September 2016.
© 2016, National Ground Water Association.
doi: 10.1111/gwat.12476
NGWA.org
one is sea water reverse osmosis (SWRO) (Ghaffour
et al. 2013). While it has a higher degree of energy
efficiency and lower operating cost compared to thermal
distillation processes, it has had continuing difficulties
with the fouling of the SWRO membranes caused by the
accumulation of biological substances on the membrane
face, which is termed biofouling (Flemming 1997;
Flemming et al. 1997; Nguyen et al. 2012). Biofouling is
caused by the complex interaction of organic substances
that precondition the surface of the membrane with a
relatively sticky coating, thereby allowing bacteria to
attach to the membrane surface and multiply to become
a biofilm. The biofilm acts to reduce the flux of sea
water through the membrane which necessitates frequent
cleaning, resulting in the reduction of the operating life
expectancy of the membrane (Nguyen et al. 2012). Algae,
bacteria, and a variety of dissolved organic compounds
are ubiquitous in sea water and must be removed before
entering the primary treatment process to reduce the rate
of biofouling.
Most SWRO systems use a direct surface water or
open-ocean intake system that necessitates the design and
operation of a complex pretreatment process train before
the feed water is permitted to pass into the primary process
(Cho et al. 2000; Missimer 2009; Valavala et al. 2011;
Sorlini et al. 2013). Elimelech and Phillip (2011) suggest
that the energy demand for pretreatment in an SWRO
plant accounts for the major part of ancillary energy usage.
However, many of the organic compounds present in
sea water are difficult to remove and still pass into the
SWRO membrane resulting in biofouling (Greenlee et al.
Groundwater
1
2009). At the macroscopic level, high concentrations of
algae during blooms (HABs) can cause the pretreatment
systems to fail, forcing facility shutdown (Caron et al.
2010; Berktay 2011; Villacorte et al. 2015). The specific
organic substances of greatest concern are bacteria,
the biopolymer (BP) fraction of natural organic matter
(NOM), and transparent exopolymer particles (TEP)
which are particulates and colloids that are produced
as extracellular excretions of algae and bacteria (BarZeev et al. 2015). TEP is a very complex and sticky
substance formed by the self-assembly of various acidic
polysaccharides and other organic compounds and has its
highest concentrations during HABs (Passow 2000, 2002;
Villacorte et al. 2009). Particulate TEP concentrations are
generally highest in shallow sea water within the photic
zone which is the depth from which most SWRO facilities
extract raw water.
The use of well systems has been demonstrated to
remove a large part of the organic substances that cause
biofouling (Rachman et al. 2014; Dehwah et al. 2015a,
2015b; Dehwah and Missimer 2016). However, little
documentation is available on the percentage removal of
each of the various types of organic matter and the various
fractions of dissolved NOM along with TEP. An unresolved issue is the impact of well age on the efficiency of
organic substance removal during aquifer transport from
the sea to the production wells. No SWRO treatment
site has been thoroughly documented that contains wells
with differing ages, so that well age can be assessed in
terms of water treatment efficiency. It is the purpose of
this research to determine whether well age impacts the
“conditioning” of a coastal aquifer in terms of removal
of particulate organic material and dissolved components
of NOM.
Methodology
Description of the Investigated Site
A SWRO plant located near Jeddah, Saudi Arabia was
selected for detailed analysis of the well intake system
(Figure 1). The North Obhor SWRO plant has been in
operation since 2001 and has undergone a number of
facility expansions in capacity. It currently has a capacity
to yield 15,350 m3 /day of treated water, which requires
43,857 m3 /day of sea water from the well intake system.
The treatment conversion efficiency of the sea water to
fresh water is about 35% because of the relatively high
salinity of the ambient sea water in the Red Sea at over
40,000 mg/L.
To obtain the necessary volume of feed water,
14 wells were constructed into a coastal, unconfined
carbonate aquifer system (Figure 2). The production
wells are similar in construction with a 30.5 cm interior
diameter, a range of total depth from 50 to 55 m and each
well contains about 36 m of screen and gravel pack. The
screened section of the production wells typically begins
at between 15 and 20 m below surface. The average
pumping rate of the wells is about 3130 m3 /day. The
wellfield is designed with an alignment parallel to the
2
A.H.A. Dehwah et al. Groundwater
shoreline with the distance between the production wells
and the mean low water ranging from 400 to 450 m. The
wells range in age from 4 months to 14 years, so the site is
ideal for assessment of well age on the degree of aquifer
treatment (Table 1). The SWRO membranes within
the facility require cleaning every 2 years on average
compared to 6 months or less for other SWRO facilities
located along the Red Sea coast that use open-ocean
intake systems (Dehwah and Missimer 2016).
Site Hydrogeology
The coastal geomorphology and shallow geology of
the Red Sea coastline of Saudi Arabia were investigated
by Dehwah et al. (2014) as part of a subsurface desalination intake feasibility assessment. In many locations along
the Red Sea coastline, a Pleistocene-age wedge of carbonate sediments was found. The unit extends from the
shelf-edge offshore at the base of the modern fringing reef
to the edge of the coastal plain or interfingers with alluvial
outwash sediments (Figure 2). The outwash sediments are
a mixture of boulders, gravel, sand, and clay which has a
generally low hydraulic conductivity except where wadi
channels exist.
Although no aquifer performance test data are
available from the site, the estimated transmissivity of
the carbonate aquifer based on well yields ranges from
2000 to 2500 m2 /day using the specific capacity estimation
technique reported by Driscoll (1986). The transmissivity
decreases from the shoreline landward based on the
reduction in aquifer thickness and hydraulic conductivity.
For comparison, the transmissivity of alluvial system is
likely to be no greater than 100 m2 /day.
Higher salinity sea water with a slightly lower pH
occurs in the carbonate sediments directly underlying the
nearshore Red Sea bottom as documented by Dehwah and
Missimer (2016). The higher salinity is likely caused by
nearshore evaporative concentration which is 2–3 m of
evaporation per year and low rates of nearshore circulation
(Dehwah et al. 2015a, 2015b).
No fresh water occurs within the alluvial sediments
landward of the carbonate aquifer and the sea water
within them typically has a slightly higher total dissolved
solids concentration compared to the sea water in the
Red Sea. The higher salinity is likely caused by past
evaporative concentration. In some nearby locations,
the alluvial sediments contain hypersaline water within
sabkha systems (Missimer et al. 2014). There is no
fresh water/sea water interface that typically occurs
within coastal aquifers. Rainfall accumulation in this
region averages less than 50 mm/year. Based on the site
hydrogeology, the pumping wells induce water flow solely
from the sea with no significant contribution from the
landward direction.
Sampling and Analysis Methodologies
Sampling
Sea water samples were collected from the sea
surface along the coast near the production wells during
NGWA.org
Figure 1. Location map.
each sample collection period to establish the raw water
concentrations of organic material before the water
entered the aquifer system. During the first sampling
campaign, 13 wells were sampled along with the reference
raw sea water sample and on a second sampling campaign
the new well (4 months in age) and a reference raw sea
water were collected (Table 1).
All water samples were collected following an established quality assurance and control protocol. The surface
sea water samples were considered to be diagnostic of the
local conditions and a water sample was collected from
the discharge of each individual production well before the
water entered and mixed in the primary intake pipeline.
The facility was in continuous operation for at least a
week before the samples were collected. Therefore, the
measured results are believed to represent true operational
conditions.
NGWA.org
The sea water samples were placed into glass bottles
and then placed into a container filled with ice to minimize
biological activity. A solution containing 0.02% (w/v) of
sodium azide was added to each total organic carbon
(TOC) sample to further reduce the potential for biological
activity. The samples were transported to the laboratory
for analysis of algae, bacteria, TOC, NOM fractions, and
TEP (particulate and colloidal) concentrations. Samples
were stored in the laboratory at a temperature of 4 ◦ C until
analyzed. The physical water quality parameters were
measured in the field.
Fundamental Physical Parameters
The collected samples were analyzed to measure the
fundamental physical water quality parameters including
turbidity, salinity, conductivity, and pH. A portable
turbidity meter (HACH 2100Q) was used to measure the
A.H.A. Dehwah et al. Groundwater
3
Figure 2. Conceptual block diagram showing wellfield with geology.
Table 1
Table 2
Sampling Points, Well Age, and Sampling Dates
Results of Analyses of Conductivity, Salinity, pH,
and Turbidity
Sampling Point Well Age
Sampling Date
Wells 1–5
14 years
Wells 6–9
11 years
Wells10–13
4 years
Well 14
4 months
Surface sea water points
SW (1)
To be used for
comparison with
wells 1 to13
SW (2)
To be used for
comparison with
well 14
November 25, 2014
November 25, 2014
November 25, 2014
February 4, 2015
November 25, 2014
February 4, 2015
sample turbidities while a portable pH meter (WTW pH
3310) was used to measure pH values. The conductivity
and salinity measurements were performed using a
portable conductivity meter (WTW Con 3210).
Microorganism Quantification
Algae and bacteria counts in the collected water samples were determined using a flow cytometer. A BD
FACSVerse flow cytometer was used to analyze the algae
cells, while an Accuri flow cytometer was used for bacterial counts. Light scattering properties and/or fluorescent
intensity was determined by the flow cytometer to distinguish between the different organism classifications (Van
der Merwe et al. 2014). Lasers were used to excite both
unstained autofluorescent organisms (algae) and stained
bacterial cells. The red laser wavelength was set at 640 nm
4
A.H.A. Dehwah et al. Groundwater
Sampling
Point
Conductivity
(ms/cm)
Salinity
(ppt)
pH
Turbidity
(NTU)
SW (1)
Well #1
Well #2
Well #3
Well #4
Well #5
Well #6
Well #7
Well #8
Well #9
Well #10
Well #11
Well #12
Well #13
SW (2)
Well #14
58.10
57.70
58.10
60.60
61.10
59.30
58.90
58.50
60.00
59.90
60.10
60.10
61.80
60.30
57.60
60.10
39.10
38.80
39.10
41.00
41.40
40.00
39.70
39.30
40.50
40.40
40.60
40.60
41.90
40.70
38.70
40.70
8.20
7.36
7.41
7.32
7.29
7.32
7.42
7.35
7.41
7.34
7.29
7.32
7.27
7.37
8.26
7.37
1.94
0.36
0.25
0.17
0.28
0.33
0.30
3.22
0.22
0.35
0.25
0.32
0.66
0.35
0.61
0.47
and the blue laser at 488 nm. Algal cell counting was
performed by combining 500 μL of each sample with a
1-μL volume of a standard containing 1-μm beads in
a 10-mL tube. The tube was then vortexed and measured using the high flow rate with a 200-μL injection
volume for 2 min. The counting procedure was repeated
three times to assess the precision of the measurements.
The different types of algae, Cyanobacteria, Prochlorococcus, and Pico/Nanoplankton, were distinguished based
on their autofluorescence as well as by the cell side angle
NGWA.org
Figure 3. TEP analysis calibration curves relating to Xanthan gum equivalent concentrations.
scatter which was used to identify them by size (Radic
et al. 2009). The detection limit for algal counts was
50 cells/mL based on the instrument calibration.
The flow cytometer method is the most accurate
means of doing counts and the precision is very high
based on the data obtained from replicate samples. While
variation occurs in the triplicate measurements, two
measurements were very close and one deviates, usually
due to large particle clogging within the water feed.
Our data showed precision of measurement ranging from
0.41% to 1.86% and averages <1%.
For bacterial counts, a comparative protocol employing SYBR®Green stain was used. A volume of 500 μL
from each sample was transferred to a 10-mL tube, incubated in a 35 ◦ C water bath for 10 min and stained with the
SYBR® Green dye (5 μL into 500-μL aliquot), vortexed,
and incubated for another 10 min. The prepared samples
were then analyzed at a medium flow setting with a 50-μL
injection volume for 1 min. Triplicate measurements were
made on each sample to assess measurement precision. The detection of the bacterial counts was about
100 cells/mL based on the instrument calibration. The precision was very good based on the replicate analysis with
typical deviation at <1%. In samples with low bacterial
concentrations close to the detection limits, the precision
can vary between 3% and 6%.
TOC and NOM Fractions
The TOC concentrations in the samples were measured using a Shimadzu TOC-VCSH instrument. The
detailed fractions of dissolved organic carbon were determined by using a Liquid Chromatography Organic Carbon
Detector (LCOCD) from DOC-Labor. The invention of
the LCOCD method and the vast improvements made in
the accuracy of the measured fractions is relatively new
and not commonly used in groundwater investigations in
the past. The protocols and methods developed by Huber
et al. (2011) were followed in order to measure different
fractions of NOM using LCOCD and have been previously described in Rachman et al. (2015) and Dehwah
and Missimer (2016).
The size exclusion chromatography column used
for this experiment was a Toyopearl HW-50S which is
produced by TOSOH. Prior to the sample measurements,
a calibration curve was established for both molecular
masses of humic substances and detector sensitivity. For
the molecular mass calibration, humic acid and fulvic
acid standards (Suwannee River Standard II) were used
NGWA.org
while potassium hydrogen phthalate and potassium nitrate
(KNO3 ) were used for sensitivity calibration (Huber et al.
2011).
The samples for the LCOCD were manually prefiltered using a 0.45-μm syringe filter to exclude the
nondissolved organics. Before analyzing the samples, a
system cleaning was performed by injection of 4000 μL
of 0.1 mol/L NaOH through the column for 260 min. Following the cleaning step, 2000 μL 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 including
BPs, humic substances, building blocks, low molecular
weight acids (LMWA), and low molecular weight neutrals
(LMWN). The manual integration was performed based
on the method developed by Huber et al. (2011). An analysis of precision conducted by Dr. Stefan Huber of DCLabor showed that the BP fraction had a precision based
on six replicates of 2.4% to 4.5%, the humic substances
ranged between 2.4% and 2.7%, the building blocks
between 3.2% and 17.5% (note that the initial concentration was very small for the sample reported), the LMWN
between 9.2% and 12.4%, and the LMWA at up to 20%
(http://www.doc-labor.de/Repro.html, accessed May 4,
2016). Based on the detections and accuracy, the data are
reported as whole numbers (Dehwah and Missimer 2016).
TEP Measurements
Two types of TEP were investigated, particulate
and colloidal. Particulate TEP has a size greater than
0.4 μm, while colloidal TEP size ranges between 0.1
and 0.4 μm (Villacorte et al. 2009). Analysis of TEP
was conducted based on the method developed by
Passow and Alldredge (1995) which includes sample
filtration, membrane staining with Alcian Blue, and
UV spectrometer measurements. A staining solution was
prepared from 0.06% (m/v) Alcian Blue 8GX (Standard
Fluka) in an acetate buffer solution (pH 4) and freshly
prefiltered through a 0.2-μm polycarbonate filter before
usage. A 300-mL volume of sea water from each
water sample was filtered through a 0.4-μm pore size
polycarbonate membrane using an adjustable vacuum
pump at low constant vacuum. After filtration, the
A.H.A. Dehwah et al. Groundwater
5
Table 3
Concentration of Algae by Type and Total, Total Bacteria, and TOC
Sampling
Point
Cyanobacteria
(cell/mL)
Prochlorococcus
(cells/mL)
Pico/Nano-Plankton
(cells/mL)
Total Algae
(cells/mL)
Bacteria
(cells/mL)
TOC
(mg/L)
SW (1)
Well #1
Well #2
Well #3
Well #4
Well #5
Well #6
Well #7
Well #8
Well #9
Well #10
Well #11
Well #12
Well #13
SW (2)
Well #14
99,420
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
76,780
140
25,455
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
7220
160
4863
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
<50
7870
<50
129,738
<100
<100
<100
<100
<100
<100
<100
<100
<100
<100
<100
<100
<100
91,870
300
520,350
9000
16,150
4250
5400
12,900
8200
63,350
8400
10,800
5050
9300
34,200
11,150
1,356,600
4300
1.1
0.4
0.4
0.3
0.3
0.3
0.4
0.8
0.4
0.5
0.3
0.4
0.3
0.3
1.1
0.3
membrane was rinsed with 10 mL of Milli-Q water to
avoid the coagulation of the Alcian Blue with possible
salt remaining on the filter after the sea water filtration.
This helps avoid overestimation of the TEP concentration.
The retained TEP particles on the membrane surface
were then stained with the Alcian Blue dye for 10 s.
After staining, the membrane was flushed with 10 mL
of Milli-Q water to remove excess dye. The flushed
membrane was then placed into a small beaker, where
it was soaked in 80% sulfuric acid for 6 h to extract the
Alcian Blue dye that was bound to the TEP. Finally, the
absorbance of the acid solution was measured using a
UV spectrometer at 752-nm wavelength to determine the
TEP concentration. The same methodology was applied to
determine the colloidal TEP. The only difference was that
a 250-mL volume of the water sample from the 0.4-μm
polycarbonate membrane permeate was filtered through a
0.1-μm pore size membrane to allow deposition of the
colloidal TEP on the membrane surface.
In order to relate the UV absorbance values to
estimated TEP concentrations, a calibration curve was
established. Xanthan gum solutions with different volumes
(0, 0.5, 1, 2, and 3 mL) were used to obtain the
calibration curves (Figure 3). The TOC concentrations
of Xanthan gum before and after 0.4-μm filtration were
analyzed, and the TOC concentration difference was
used to calculate the gum mass on each filter and the
TEP concentration was estimated using the calibration
curve. The same procedures were used for the 0.1-μm
membrane to establish the calibration curve for colloidal
particles. Afterward, the TEP concentration was expressed
in terms of Xanthan gum equivalent in μg Xeq./L by
dividing the TEP mass on the corresponding volume of
TEP samples. Because particulate and colloidal TEP are
determined indirectly, these values must be considered to
be semiquantitative.
6
A.H.A. Dehwah et al. Groundwater
Results
Physical Water Parameters
The physical water quality parameters consisted of
conductivity, salinity, pH, and turbidity. As shown in
Table 2, the conductivity and salinity of the sea water were
slightly higher in the wells compared to the surface sea
water in most cases. The pH of the surface sea water was
significantly higher than the well water with a comparison
of about 8.2 to 7.3, respectively. The turbidity of the well
discharge was normally significantly lower than surface
sea water with the exception of well #7.
Algae and Bacteria Concentrations
Three different types of algae were detected
in the surface sea water, including Cyanobacteria,
Prochlorococcus, and Pico/Nanoplankton (Table 3). The
Cyanobacteria were the major contributor to the overall
algal population. Concentrations of algae between the
surface sea water and the well discharges showed nearly
a complete elimination during aquifer transport. The
algae concentrations were below the significant detection
limits of the analysis method at 100 cells/mL for the
total counts and 50 cells/mL for the fractions except
in a single well. The only algae counts in the wells
above the detection limits occurred in well #14. While
these concentrations were significant, they represented
a very small fraction of those occurring in the surface
sea water.
The concentration of bacteria showed a very large
difference between that in the surface sea water and in
the well discharges (Table 3). It appears that a large part
of the bacteria concentration was removed in the aquifer
during transport from the sea to the wellheads. Wells #7
and #12 showed higher concentrations compared to the
other production wells.
NGWA.org
Figure 4. Natural organic matter fraction (μg/L) concentration comparisons between the surface sea water and the wells.
Figure 5. Particulate and colloidal TEP concentrations of Xanthan gum equivalent (μg Xeq./L).
TOC and NOM Fraction Concentrations
Some of the TOC was removed during aquifer
transport from the sea to the production wells in all
cases. Most commonly, the TOC concentration was
reduced during transport between 64% and 73% (Table 3).
However, wells #7 and #9 showed a lower removal
percentage at 27.3 and 54.5%, respectively.
The NOM fractions represent the compositional
breakdown of dissolved organic carbon. The fractions
are shown in Figure 4 from BPs at the base with
humic substances (HS), building blocks (BB), LMWN,
and LMWA, which is the general order of overall
molecular weight (from the highest to the lowest). The
smallest fraction concentration in the background sea
water samples is the low molecular weight acid fraction
followed by the BP fraction. The humic substances
fraction had the highest concentration followed by the
LMWN during the first sampling, but this was reversed
during the second sampling. The building blocks were in
the middle of the concentration order.
The measured NOM fractions in the wells show that
the average of all concentrations has a pattern of high to
low concentration of LMWN, humic substances, building
blocks, LMWA, and BPs. The average of the reference sea
water samples shows the same order of concentration at
the high end, but the BPs have a higher concentration
NGWA.org
compared with the LMWA. It must be noted that the
LMWN concentrations were actually higher in some of the
well discharges compared to the surface sea water samples
(wells #1, #7, and #13). The LMWA concentrations in
some of the well discharges were also significantly greater
than those in the background sea water samples (wells #1,
#7, and #14).
Particulate and Colloidal TEP Concentrations
In all cases, the particulate and colloidal TEP
concentrations found in the well discharges were less than
those in the reference sea water samples (Figure 5). In
the surface sea water samples, the particulate TEP had
a greater concentration compared to the colloidal TEP
which is typical based on the fact that the colloidal TEP
material is generally considered to be the precursor to the
formation of the particulate TEP. There was a very large
range of TEP concentrations in the well discharges.
Discussion
Physical Parameter Changes During Aquifer Transport
During transport from the Red Sea to the wells, the
conductivity and the salinity increased slightly and the pH
decreased slightly, and in general, the turbidity reduced
with one exception. This same pattern of increased salinity
A.H.A. Dehwah et al. Groundwater
7
and decreased pH has been found beneath the seabed in
this regional and is not unusual (Dehwah and Missimer
2016). The production wells are located in a zone along
the shoreline at distances ranging between 400 and 450 m
(Figure 2), but during pumping the wells induce water
movement from both the seaward and landward sides
of the site. The hydraulic conductivity to the seaward
side is greater within the carbonate aquifer, so that is
the predominant direction of recharge. Commonly, the
alluvial aquifer system located landward can contain
higher salinity water which also may cause the well
discharge water to have a slightly higher salinity, but
the similar occurrence of higher salinity and lower pH
immediately beneath the seabed is the most likely cause
of this difference.
The transport of sea water from the shallow nearshore
to the production wells is clearly demonstrated by the
occurrence of shallow water algae, Cyanobacteria and
Prochlorococcus in one of the production wells and the
presence of particulate TEP. It also should be noted that
the major cation and anion ratios in the well water are
identical to that in the nearshore Red Sea water.
The lower pH of well water is likely caused by
calcium carbonate precipitation between the sea water
entering the aquifer and during its travel to the production
wells. There are modern marine hardgrounds covering the
shallow seafloor at the site. These deposits are cemented
with acicular aragonite and high magnesium calcite and
occur in the inter-reef area of most of the Red Sea
offshore.
Filtration occurs as sea water passes through the
carbonate aquifer and into the wells. This filtration reduces
the concentration of suspended sediments that produce
turbidity by size exclusion. The higher turbidity measured
in well #7 discharge may be indicative of some aquifer
material than that could have been incorporated into the
discharge during pumping.
Aquifer Treatment and the Reduction of Organic Carbon
During Transport
The pumping-induced transport of sea water through
the aquifer and into the production wells was found to
be very effective in the removal of the most critical
organic materials. The compounds, in terms of biofouling
reduction, were compared to a reference surface sea
water sample collected at the same time as the well
water samples were collected (Table 4). There was some
seasonal variability in the organic parameters measured
and the reference samples may not be representative of all
conditions that could occur. While there is some question
concerning the use of the reference samples in terms
of comparative organic matter analyses and chemistry to
the well discharges, because water entering the aquifer
from the sea is not uniform and can occur in a belt
covering some unknown portion of the nearshore marine
environment, not a single point. However, the surface
water sample is representative of the water quality that
would occur if a typical surface water intake system was
used to supply the SWRO plant with raw water. Therefore,
8
A.H.A. Dehwah et al. Groundwater
the aspect concerning “removal” is qualified in terms of
application to the SWRO operation.
The largest particulate organic materials are the algae
and bacteria which were removed during transport at
percentages greater than 99.7 and 93.2% respectively. The
TOC was removed at greater than 64% in all cases with
the exception of groundwater from two wells.
Removal of algal and bacteria during transport is
likely caused primarily by straining as the organisms pass
into and through the aquifer. Also, the organisms may be
adsorbed to the aquifer matrix which becomes coated with
NOM, particularly with sticky polysaccharides within the
BP fraction of NOM and TEP. The bacteria are removed
by straining, adsorption, and perhaps by predation of other
bacteria in the aquifer. It is not known if the bacteria
counted are all from the sea or some are indigenous to
the aquifer system containing sea water.
The substances of perhaps the greatest concern that
contribute to biofouling are the BPs and both particulate
and colloidal TEP (Baek et al. 2011; Filloux et al.
2012; Bar-Zeev et al. 2015). Aquifer treatment shows a
reduction of between 90.9% and 100% of the BP fraction
of NOM. The aquifer treatment of TEP was less effective
with removal of the particulate fraction ranging from
36.2% to 100% and the colloidal fraction from 16.7% to
91.1% with the average removal percentages being 75.7%
and 56.4%, respectively.
Other NOM fractions showed major reductions in
concentration, particularly the humic substances and
building blocks. The range in concentration reductions
was 50.7% to 78.0% in the humic substances and 22.1% to
71.6% in the building blocks with corresponding average
reductions of 65.6% and 51.6%, respectively.
The LMWN and LMWA concentrations differ vastly
during aquifer transport with both increases and decreases
in concentration. The range of concentrations of the
LMWN showed a range of plus 43.6% to minus 42.3%
with an average of a 12.1% reduction. The range of
concentrations of the LMWA showed a range of plus
233.3% to a reduction of 61.1% with an average increase
of 3.9%. These differences can be expected because
during aquifer transport, the larger molecular weight
fractions can be biochemically broken down into smaller
molecular weight substances represented by these fraction
groups.
It is well known that carbonate aquifer systems
provide a high degree of TOC and DOC removal (Suess
1970; McConnell and Hacke 1993; Jin and Zimmerman,
2010). Suess (1970) reported that 14% of TOC was
adsorbed and Jin and Zimmerman (2010) suggested that
abiotic interactions between NOM and the carbonate
aquifer matrix is a substantial controlling factor for
removal of NOM. Carbonate aquifers are also particularly
effective in the adsorption of fatty acids and carboxylated
polymers (Thomas et al. 1993). Alginic acid, which can
be detected using a polysaccharide stain (Alcian Blue),
is removed at high percentages (Perry et al. 2004). This
group of substances does have similarity to TEP in sea
water which is produced by extracellular excretion.
NGWA.org
Table 4
Comparison of Various Organic Particles and Substances with Percentage of Removal by Aquifer Treatment
based on Reference Sample Comparison
Wells
Well Age
Algae
Bacteria
TOC
BP
HS
BB
LMWN LMWA
TEP-P
TEP-C
(Years) (cell/mL) (cells/mL) (mg/L) (μg/L) (μg/L) (μg/L) (μg/L)
(μg/L) (μg Xeq./L) (μg Xeq./L)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
14
14
14
14
14
11
11
11
11
4
4
4
4
0.33
>99.9
100
>99.9
100
100
100
100
100
100
100
100
100
100
100
98
97
99
99
97
98
88
98
98
99
98
93
98
100
64
64
73
73
73
64
27
64
54
73
64
73
73
73
96
91
100
96
96
96
96
91
96
98
96
98
93
91
58
52
78
78
72
56
51
54
56
74
71
78
68
73
30
45
69
70
59
43
22
46
44
54
62
72
60
45
+371
12
37
42
32
5
+43
32
9
27
32
29
+44
35
+72
28
56
50
50
28
+233
50
61
+39
39
44
6
+121
28
100
92
97
75
93
36
87
67
81
84
36
79
44
27
76
91
76
41
52
78
44
61
36
50
17
66
76
1 (+) indicates that measured value at the well was higher than the corresponding value of the reference sea water sample.
The aquifer matrix material, limestone or siliciclastic,
does not appear to have a distinct impact on the removal
percentage of the NOM fractions based on a number of
investigations documented in Rachman et al. (2015). They
also concluded that the length of the travel path from the
sea to the production wells and the residence time of the
infiltrated sea water in the aquifer may be the dominant
factors controlling the rate of reduction. However, the
removal of TEP does seem to occur at higher percentages
in carbonate compared to siliciclastic aquifers (Rachman
et al. 2015), which confirms what was observed by Perry
et al. (2004).
There is some evidence found in the data collected
herein that showed the biochemical removal of the heavier
fractions of NOM (e.g., BPs and humic substances). Some
increases in the concentrations of the LMWA and neutrals
suggest that removal mechanism is active in the aquifer.
The key finding within the NOM fractions analyses
is that aquifer treatment is very effective at reducing
the concentrations of the most important substances that
have considerable impact on the biofouling of SWRO
membranes, which are the BPs and part of the humic
substances (Filloux et al. 2012). These fractions are
believed to contain most of the TEP substances. Also,
TEP is significantly reduced in concentration and in many
cases to a lower degree than occurs in some engineered
pretreatment designs (Dehwah and Missimer 2016). The
carbonate aquifer system investigated does seem to reduce
the reduction percentage of TEP.
Impact of Well Age on Removal of Organic Material
A comparison of the average values for the various
organic constituents shows that well age does not have
a significant impact on the removal percentage during
aquifer transport (Table 5). The only substance that could
be used to support the age having significance was the
BP fraction in the groundwater collected from the most
NGWA.org
Table 5
Comparison of Impact of Well Age on Organic
Substances Removal Efficiencies
Organic Types
14 Years 11 Years 4 Years 4 Months
Algae (cell/mL)
Bacteria (cells/mL)
TOC (mg/L)
Biopolymers
(μg/L)
Humic substances
(μg/L)
Building blocks
(μg/L)
Low molecular
weight neutrals
(μg/L)
Low molecular
weight acids
(μg/L)
TEP—particulate
(μg Xeq./L)
TEP—colloidal
(μg Xeq./L)
>99.9
98.2
69.1
96.1
>99.9
95.6
52.3
95.2
100
97.1
70.5
96.5
99.7
99.7
72.7
90.9
67.6
54.3
72.6
72.9
54.9
38.2
61.8
44.9
17.3
0.8
11.2
34.9
22.22
+23.61
12.5
+121.4
89.7
71.0
70.0
43.7
62.2
58.9
41.9
75.8
1 (+) indicates that measured value at the well was higher than the
corresponding value of the reference sea water sample.
recently installed well (#14) which has only a 90%
removal rate. However, it is concluded that well age does
not exhibit any greater impact on the aquifer treatment
process at a constant distance of transport between 400
and 450 m. This general conclusion may not be applicable
to other locations where the distance of transport is shorter
or the aquifer contains significant preferable transport
corridors, such as in mature karst aquifers.
The aquifer investigated is a heterogeneous carbonate system which is likely to have influenced some of
A.H.A. Dehwah et al. Groundwater
9
the variation in the removal of some of the organic
compounds, such as TEP. The bacterial processes acting
within the aquifer are likely randomly distributed and,
thereby variation in produced water organic concentrations was observed. The variation in distance of 400 to
450 m between the wells and the sea water source does
not appear to have any influence on the organic concentrations measured.
Conclusions
Aquifer treatment of sea water has been demonstrated
to provide a significant degree of pretreatment for removal
of particulate and dissolved organic matter. Reference
samples of nearshore surface sea water were used to
make assessments of the removal during aquifer transport.
It is known that these reference samples may not be
representative of the sea water quality entering the aquifer
to replace pumped water from the wells in space and
time. However, if a conventional surface intake system
was used to supply the SWRO plant with raw water, the
samples would be representative of the water that would
enter the facility. Therefore, the “removal” percentages are
valid from a facility operational perspective and that the
aquifer is indeed providing a major pretreatment function.
Nearly all of the algae, over 90% of the bacteria,
a significant percentage of the fractions of NOM, and a
significant percentage of particulate and colloidal TEP are
removed during aquifer transport. The aquifer matrix, in
this case carbonate, does not seem to have a material effect
on the removal of NOM fractions in general, but a higher
percentage of TEP removal likely occurs in carbonate
systems based on the data obtained at this site compared to
the literature. Removal of these substances has a material
effect on SWRO desalination of sea water in that it
reduces the rate of membrane biofouling, reduces the
degree of engineered pretreatment processes that have to
be operated, and reduces the overall cost of desalination.
An analysis of the effect of well age on the degree
of organic matter reduction achieved showed that it is not
very significant. The aquifer system provides a very rapid
and robust degree of treatment by straining, absorption,
adsorption, and biochemical degradation of NOM. It
appears that long-term “conditioning” of the aquifer is not
required to provide a high degree of sea water treatment.
At a distance of 400 to 450 m between the sea and
the production wells, a similar degree of treatment was
achieved in wells ranging from 4 months to 14 years of
operation. Variations in the degree of treatment may be
caused by heterogeneity with the carbonate aquifer into
which the wells were constructed and by variation in
biochemical activity with the aquifer.
Removal of the various fractions of NOM appears
to be based on the molecular weight with the highest
being removed at the greatest percentage. The order of
removal from high to low is best demonstrated from
the BPs to the humic substances to the building blocks.
However, the LMWN and LMWA sometimes increased
in concentration which may be caused by biochemical
10
A.H.A. Dehwah et al. Groundwater
breakdown of the larger organic molecules into smaller
molecular weight compounds during aquifer transport and
biochemical breakdown.
Acknowledgments
Funding for this research was provided by King
Abdullah University of Science and Technology, Thuwal,
Saudi Arabia. The authors thank the Water Desalination
and Reuse Center for the use of analytical equipment. The
authors would like to thank Khaled Bin Bandar for field
support, SAWACO company team, and Engineers Nizar
Kammourie, Najm El-Jafery, and Firas Yaish for access
to the facilities and on-site support.
References
Baek, Y., J. Yu, S.-H. Kim, S. Lee, and J. Yoon. 2011. Effects
of surface properties of reverse osmosis membranes on
biofouling occurrence under filtration conditions. Journal
of Membrane Science 382: 91–99.
Bar-Zeev, E., U. Passow, S.R.-V. Castrillon, and M. Elimelech.
2015. Transparent exopolymer particles: From aquatic environments and engineered systems to membrane biofouling.
Environmental Science & Technology 49: 691–707.
Berktay, A. 2011. Environmental approach and influence
of red tide to desalination process in the middle east
region. International Journal of Chemical & Environmental
Engineering 2, no. 3: 183–188.
Caron, D.A., M.-E. Garneau, E. Seubert, M.D.A. Howard, L.
Darjany, A. Schnetzer, I. Cetinic, G. Filteau, P. Lauri,
B. Jones, and S. Trussel. 2010. Harmful algae and their
potential impacts on desalination operations off southern
California. Water Research 44: 385–416.
Cho, J., G. Amy, and J. Pellegrino. 2000. Membrane filtration
of natural organic matter initial comparison of rejection
and flux decline characteristics with ultrafiltration and
nanofiltration membranes. Water Research 33, no. 11:
2517–2526.
Dehwah, A.H.A., and T.M. Missimer. 2016. Subsurface intake
systems: Green choice for improving feed water quality
at SWRO desalination plants, Jeddah, Saudi Arabia. Water
Research 88: 216–224.
Dehwah, A.H.A., K.Z. Jadoon, S. Al-Mashhaarwi, and T.M.
Missimer. 2015a. Effects of nearshore-evaporation rates
on the design of seabed gallery intake systems for
SWRO facilities located along the Red Sea shoreline
of Saudi Arabia. Desalination and Water Treatment.
DOI:10.1080/19443994.2015.1098796.
Dehwah, A.H.A., S. Li, S. Al-Mashharawi, H. Winters, and
T.M. Missimer. 2015b. Changes in feedwater organic
matter concentrations based on intake type and pretreatment
processes at SWRO facilities, Red Sea, Saudi Arabia.
Desalination 360: 19–27.
Dehwah, A.H.A., S. Al-Mashhawari, and T.M. Missimer. 2014.
Mapping to assess feasibility of using subsurface intakes
for SWRO, Red Sea coast of Saudi Arabia. Desalination
and Water Treatment 52: 2351–2361.
Drewes, J.E., M. Reinhart, and P. Fox. 2003. Comparing
microfiltration-reverse osmosis and soil-aquifer treatment
for indirect potable reuse. Water Research 37, no. 15:
3612–3621.
Driscoll, F.G. 1986. Groundwater and Wells, 2nd ed. St. Paul,
Minnesota: Johnson Division.
Elimelech, M., and W.A. Phillip. 2011. The future of seawater
desalination: Energy, technology, and the environment.
Science 333, no. 6043: 712–717.
NGWA.org
Filloux, E., H. Gallard, and J.-P. Croue. 2012. Identification of effluent organic matter fractions responsible
for low-pressure membrane fouling. Water Research 46:
5531–5540.
Flemming, H.-C. 1997. Reverse osmosis membrane biofouling.
Experimental Fluid Science 14: 382–391.
Flemming, H.-C., G. Schaule, T. Griebe, J. Schmitt, and A.
Tamachkiarowa. 1997. Biofouling—the Achilles heel of
membrane processes. Desalination 113, no. 2: 215–225.
Ghaffour, N., T.M. Missimer, and G.L. Amy. 2013. Technical
review and evaluation of the economics of desalination:
Current and future challenges for better sustainability.
Desalination 309: 197–207.
Greenlee, L.F., D.F. Lawler, B.D. Freeman, B. Marrot, and P.
Moulin. 2009. Reverse osmosis desalination: Water sources,
technology, and today’s challenges. Water Research 43, no.
9: 2317–2348.
Hubbs, S.A. 2005. Riverbank Filtration Hydrology-impacts on
System Capacity and Water quality. Dordrecht, Netherlands: Springer.
Huber, S.A., A. Balz, M. Albert, and W. Pronk. 2011.
Characterization of aquatic humic and non-humic matter with size-exclusion chromatography—organic carbon
detection—organic nitrogen detection (LC-OCN-OND).
Water Research 45, no. 2: 879–885.
Jin, J., and A.R. Zimmerman. 2010. Abiotic interactions of
natural dissolved organic matter and carbonate aquifer rock.
Applied Geochemistry 25: 472–484.
McConnell, J. B., and C. M. Hacke. 1993. Hydrogeology, water
quality, and water-resources development potential of the
Upper Floridan aquifer in the Valdosta aerea, south-central
Georgia. U. S. Geological Survey Science Investigation
Report 93-4044, Reston, Virginia: USGS, 44 pp.
Missimer, T.M. 2009. Water supply development, aquifer storage, and concentrate disposal for membrane water treatment facilities, 2nd ed. Sugar Land, Texas: Schlumberger
Limited.
Missimer, T.M., K.Z. Jadoon, D. Li, C. Hoppe-Jones, and S.
Al-Mashharawi. 2014. Hydrogeology and water quality of
a coastal alluvial aquifer and its potential use as an intake
system for a seawater reverse osmosis water treatment
system, Thuwal, Saudi Arabia. Hydrogeology Journal 22,
no. 8: 1921–1934.
Missimer, T.M., N. Ghaffour, A.H.A. Dehwah, R.M. Rachman,
R.G. Maliva, and G.L. Amy. 2013. Subsurface intakes
for seawater reverse osmosis facilities: Capacity limitation,
water quality improvement, and economics. Desalination
322: 37–51.
Nguyen, T., F.A. Roddick, and L. Fan. 2012. Biofouling of
water treatment membranes: A review of underlying causes,
monitoring techniques and control measures. Membranes 2:
802–840.
Passow, U. 2002. Production of transparent exopolymer particles
(TEP) by phytoplankton and bacterioplankton. Marine
Ecology Progress Series 236: 1–12.
Passow, U. 2000. Formation of transparent exopolymer particles,
TEP, from dissolved precursor material. Marine Ecology
Progress Series 192: 1–11.
NGWA.org
Passow, U., and A.L. Alldredge. 1995. A dye-binding assay
for the spectrophotometric measurement of transparent
exopolymer particles (TEP). Limnology & Oceanography
40: 1326–1335.
Perry, T.D. IV, O.W. Duckworth, C.J. McNamara, S.T. Martin,
and R. Mitchell. 2004. Effects of the biologically produced
polymer alginic acid on macroscopic and microscopic calcite dissolution rates. Environmental Science & Technology
38: 3040–3046.
Rachman, R.M., A.H.A. Dehwah, S. Li, H. Winters, S. AlMashharawi, and T.M. Missimer. 2015. Chapter 9: Effects
of well intake systems on removal of algae, bacteria, and
natural organic matter. In Intakes and Outfalls for Seawater
Reverse-Osmosis Desalination Facilities: Innovations and
Environmental Impacts, ed. T.M. Missimer, B. Jones,
and R.G. Maliva, 163–193. Basal, Switzerland: Springer
International Publishing.
Rachman, R.M., S. Li, and T.M. Missimer. 2014. SWRO
feed water quality improvement using subsurface intakes
in Oman, Turks and Caicos Islands, and Saudi Arabia.
Desalination 351: 88–100.
Radic, T., T. Silovic, D. Santic, D. Fuks, and M. Micic. 2009.
Preliminary flow cytometric analysis of phototrophic picoand nanoplankton communities in the Northern Adriatic.
Fresenius Environmental Bulletin 18: 715–724.
Ray, C., G. Melin, and R.B. Linsky (Eds). 2002. Riverbank
Filtration. Fountain Valley, California: National Water
Research Institute.
Sorlini, S., F. Gialdini, and C. Collivignarelli. 2013. Removal
of cyanobacteria cells and Microcystin-LR from drinking
water using a hollow fiber microfiltration pilot plant.
Desalination 309: 106–112.
Suess, E. 1970. Interaction of organic compounds with calcium carbonate-I: Association phenomena and geochemical
implications. Geochimica et Cosmochimica Acta 34, no. 2:
157–168.
Thomas, M.M., J.A. Clouse, and J.M. Clouse. 1993. Adsorption
of organic-compounds on carbonate minerals. 3. Influence
on dissolution rates. Chemical Geology 109: 227–237.
Valavala, R., J. Sohn, J. Han, N. Her, and Y. Yoon. 2011.
Pretreatment in reverse osmosis seawater desalination: A
short review. Environmental Engineering Research 16, no.
4: 205–212.
Van der Merwe, R., F. Hammes, S. Lattemann, and G.L.
Amy. 2014. Flow cytometric assessment of microbial
abundance in the near-filed area of seawater reverse osmosis
concentrate discharge. Desalination 343: 208–216.
Villacorte, L.O., S.A.A. Tabatabai, D.M. Anderson, G.L. Amy,
J.C. Schippers, and M.D. Kennedy. 2015. Seawater reverse
osmosis desalination and (harmful) algae blooms. Desalination 360: 61–80.
Villacorte, L.O., M.D. Kennedy, G.L. Amy, and J.C. Schippers.
2009. Measuring transparent exopolymer particles (TEP) as
indicator of the (bio) fouling potential of RO feed water.
Desalination and Water Treatment 5: 207–212.
A.H.A. Dehwah et al. Groundwater
11