Desalination 203 (2007) 243–255

Water quality and performance evaluation at seawater reverse

osmosis plants through the use of advanced analytical tools

Jérôme Leparca*, Sophie Rapennea,b, Claude Courtiesc, Philippe Lebaronc,

Jean Philippe Crouéb, Valérie Jacquemeta, Greg Turnerd
a
Anjou Recherche – Veolia Water, Chemin de la Digue, 78600 Maisons Laffitte, France
Tel. +33 (1) 34 93 31 33; Fax +33 (1) 34 93 31 10; email: [email protected]
b
LCEE/ESIP UMR CNRS 6008, Université de Poitiers, 40 avenue du recteur Pineau, 86022 Poitiers, France
c
Observatoire Océanologique de Banyuls, Université Pierre et Marie Curie, UMR CNRS No. 7621 and 7628,
Rue Fontaulé, 66651 Banyuls-sur-Mer, France
d
Veolia Water Industrial Outsourcing, Blackwell House, 1 Three Valley Way, Bushey, Herts, United Kingdom

Received 2 February 2006; accepted 28 March 2006

Abstract
Advanced analytical tools have been developed to allow thorough characterisation of seawater samples from
many aspects: (i) inorganic characterisation, (ii) characterisation of the natural organic matter, (iii) quantification
of the main algal pigments (chlorophyll, pheophytin), and (iv) enumeration of phytoplankton and bacteria. These
analytical tools were used at a seawater reverse osmosis plant in Gibraltar and provided valuable information to
quantify the risks of fouling. Indeed, in addition to the measurement of the Silt Density Index (SDI), the results of
the advanced analyses provided a detailed characterisation of the raw seawater sources (surface seawater and well
seawater) as well as a thorough assessment of the pretreatment performance.
Keywords: Reverse osmosis; Desalination; Seawater quality; Natural organic matter; Phytoplankton; Bacteria

*Corresponding author.

Presented at EuroMed 2006 conference on Desalination Strategies in South Mediterranean Countries: Cooperation
between Mediterranean Countries of Europe and the Southern Rim of the Mediterranean. Sponsored by the
European Desalination Society and the University of Montpellier II, Montpellier, France, 21–25 May 2006.

0011-9164/07/$– See front matter © 2007 Published by Elsevier B.V.

doi:10.1016/j.desal.2006.03.529
244 J. Leparc et al. / Desalination 203 (2007) 243–255

1. Introduction Gibraltar. The Design Build Own and Operate

contract was awarded for a 10 year period and
Design and operation of seawater reverse os-
allowed for a complex integration of the new plant
mosis plants strongly depend on the raw seawater
with a capacity of 1400 m3/d.
quality to be treated. The performance of desalina-
tion reverse osmosis (RO) systems relies upon the
2.1. Raw water sources
production of high quality pretreated water. To
date, membrane manufacturers have put a lot em- The SWRO plant is fed by two sources:
phasis on the Silt Density Index (SDI) as a surro- • an open seawater intake,
gate parameter for water quality to prevent fouling • and a rock-well (seawater well).
on RO membranes. However, as already stated
Under normal operating conditions, the plant
by numerous researchers [1,2], the SDI measure-
feedwater is composed about of 2/3 of well sea-
ment has several limitations for reverse osmosis
water and 1/3 of seawater from the open intake.
plants. Indeed, this measurement is based on the
reduction of permeability with time of a pretreated
water sample through a microfiltration (MF) 2.2. Description of the pretreatment process
membrane. The transfer mechanisms through a Fig. 1 depicts the flow diagram of the pretreat-
RO membrane being very different from those ment process upstream of the two RO trains. The
occurring through a MF membrane, it is highly raw water from both sources are combined at the
likely that the fouling mechanisms and occur- outlet of the rock-well, and the combined raw
rences will greatly differ from a SDI membrane water feeds a holding tank, which provides a 2–
and a RO membrane. Therefore, whilst the SDI 3 h contact time for removal of the larger debris
measurement can be a useful indicator with respect (e.g. sand). Then, two chemicals are added just
to the particulate content and MF fouling com- upstream of the three pressurized dual media
pounds, this index does not provide any informa- filters (DMF):
tion regarding the nature of the foulants passing • the cationic polymer Genefloc GPF (aqueous
through a 0.45 µm membrane and the risks of bio- solution of polyquaternary amine; supplier:
fouling. Thereby, other water quality indicators Genesys) — dose of 2 mg/L as commercial
must be developed to assess the quality of pre- product
treated seawaters feeding desalination reverse os- • sodium hypochlorite — dose of 0.3 mg/L
mosis systems. To that end, Veolia Water has de-
Free chlorine residuals measured in both the
veloped various analytical tools in order to provide
influent and the effluent of the filters are as fol-
a better characterisation of raw seawater samples
lows:
as well as a better performance assessment of
• free chlorine residual upstream of the filters
pretreatment processes. This paper presents the
ranged between 0.07 and 0.15 mg/L Cl2
application of these tools for assessing the per-
• no free chlorine residual was detected down-
formance of the Glen Rocky seawater reverse
stream of the filters
osmosis (SWRO) plant located in Gibraltar.
The three dual media filters (see Fig. 2) have
the following characteristics:
2. Presentation of the Glen Rocky SWRO Plant
• hydraulic loading rate of 11 m/h when all filters
The British Ministry of Defence selected on duty and 16 m/h when one filter is being
Veolia Water Industrial Outsourcing (VWIO) for backwashed (two filters on duty)
the design, construction and operation of a sea- • 90 cm of sand
water reverse osmosis plant on the peninsular of • 30 cm of anthracite
 J. Leparc et al. / Desalination 203 (2007) 243–255 245

Flocculant

Seawater Intake Genefloc GPF Antiscalant

DMF Genesys LF To cartridge filter and RO
Dosage = 2 mg/L
Lane B

Seawater DMF
holding tank

NaOCl NaHSO3
DMF To cartridge filter and RO
0.3 mg/L Lane A
Seawater Rockwell
Measured Measured
free chlorine free chlorine
= 0.07 – 0.15 = 0 mg/L
mg/L

Fig. 1. Pretreatment process.

• 1 pressure vessel containing 26 cartridge filters

(pore size: 5 µm)
• 1 high pressure pump + 2 energy recovery
devices (pressure exchangers from ERI — see
picture)
• 8 pressure vessels in parallel, each containing
7 RO modules (Dow Filmtec SWHR 380)
• 1 cleaning chemical tank with a recirculation
pump
Both lanes have a recovery ratio of 39% and
the permeate flux ranges between 14 and 15 L/m2-h.
Fig. 3 shows a photograph of the two energy re-
covery devices on a single RO lane.
Fig. 2. Photograph of the three pressurized dual media
filters operated in parallel.

2.4. Performance of the RO units at the Glen

Sodium bisulphite and an antiscalant (Genesys Rocky desalination plant
LF; dose = 1.8 ppm) are dosed upstream of the
Since start-up in September 2003, only three
two reverse osmosis units for chlorine removal
chemical cleanings have been performed. Also, it
and scaling control, respectively.
was found that some of the periods with increased
differential pressure were due to mechanical issue
2.3. Description of the RO units
or faulty pressure transmitters and not due to
Downstream of the three pressurized filters, membrane fouling. Figs. 4 and 5 respectively de-
two similar reverse osmosis skids treat the filtered pict the normalized flow and differential pressure
water. Each RO lane is composed of: computed as recommended by Dow Filmtec [3].
246 J. Leparc et al. / Desalination 203 (2007) 243–255

2.5. Water quality parameters

Table 1 summarizes the analyses performed
for the seawater characterization and the perform-
ance assessment of the pretreatment process at the
Glen Rocky SWRO plant.

2.6. LC–OCD chromatography or size exclusion

chromatography
The LC-OCD system (liquid chromatography-
organic carbon detection) consists of a size ex-
clusion chromatography column, which separates
hydrophilic organic molecules according to their
molecular size. The underlying principle is the
diffusion of molecules into the resin pores. This
means that larger molecules elute first as they can
not penetrate the pores very deeply, while smaller
molecules take more time to diffuse into the pores
and out again. The separated compounds are then
detected by two different detectors: a UV detector
(absorption at 254 nm) and a DOC detector (after
inorganic carbon purging). Depending on the size
of the molecules, the composition of the organic
Fig. 3. Photograph of the ERI pressure exchangers. matter can be obtained.

Fig. 4. Normalized permeate flow.

 J. Leparc et al. / Desalination 203 (2007) 243–255 247

Fig. 5. Normalized differential pressure.

Table 1 With a bespoke algorithm program, the dif-

Analyses and analytical methods ferent peaks can be integrated to evaluate the pro-
portion of each organic fraction. The total organic
Parameter Analytical method
carbon measurement can be carried out in the by
Cations (sodium, ICP – optical emission
pass mode. In this case, the samples go straight
potassium, calcium, spectroscopy
magnesium, strontium, through the TOC reactor and analyser.
barium)
Boron Spectrophotometry
Chloride Potentiometry 3. Raw water quality
Bromide Ionic chromatography Table 2 summarizes the raw seawater quality
Hydrogenocarbonate Alkalimetric titration
data collected between June 21 and June 24, 2005.
Sulphate Ionic chromatography
Silica Colorimetry Temperature, pH, turbidity, and conductivity were
Fluoride Ionometry recorded on each sample. SDI measurements were
Phosphate, nitrate, nitrite, Continuous flow performed according to the ASTM method [4].
silicate The SDIs were computed at three different times
TOC and total nitrogenShimatzu TOC-V analyzer (3, 5, and 15 min) according to the following equa-
(high temperature catalytic
tion:
oxidation)
UV254 UV spectrometry
Amino acids HPLC 100 ⎛ t ⎞
Polysaccharides Spectrophotometry
SDI X = × ⎜1 − 0 ⎟
X ⎝ tX ⎠
Monosaccharides Spectrophotometry
LC-OCD chromatography See details blow
Picophytoplankton Flow cytometry where X total elapsed flow time, min (3, 5, or
Bacteria Flow cytometry 15 min); t0 initial time required to collect 500 mL of
Chlorophyll and Fluorimetry sample, s; tX time required to collect 500 mL of
pheophytin sample after test time X (3, 5, or 15 min), s.
248 J. Leparc et al. / Desalination 203 (2007) 243–255

Table 2
Raw water quality parameters (samples collected between June 21 and June 24, 2005)

Combined raw water1 Well seawater Surface seawater

(open intake)
Temperature, °C 18.6–20.5 20.1 19.3
pH 7.87–7.92 7.80 8.07
Turbidity, NTU 0.4–1.1 0.3 1.6
Conductivity, mS/cm @ 20°C 49.9 49.9 49.6
Na+, mg/L 10,841 10,747 10,945
K+, mg/L 402 399 410
Ca2+, mg/L 441 438 441
Mg2+, mg/L 1,355 1,340 1,371
Ba2+, µg/L 15 20 10
Sr2+, mg/L 6.5 6.4 6.7
Cl–, mg/L 20,700 20,100 20,900
Br–, mg/L 69 68 69
HCO–3, mg/L 162 166 161
SO2–4 , mg/L 2,855 2,840 2,965
F–, mg/L 1.35 1.35 1.55
Phosphate, mg/L 0.05 0.09 0.02
Nitrate, mg/L 1.1 1.9 0.2
Nitrite, µg/L 4 1 6
Silicate, mg/L 1.4 2.4 0.4
Boron 4.9 4.8 5.1
Silica, mg/L 0.9 1.5 0.2
TOC, mg/L 0.65 0.57 0.80
Total nitrogen, mg/L 0.26 0.41 0.11
Total amino acids, mg/L 0.01 0.01 0.01
UV254, /m 0.68 0.72 0.70
Polysaccharides, mg/L 0.31 0.17 0.20
Monosaccharides, mg/L 0.09 0.09 0.10
Picophytoplankton
Total picophytoplankton, nb./mL 1289 181 3718
Synechococcus, nb./mL 754 91 2036
Picoeucaryote, nb./mL 473 84 1477
Nanoeucaryote, nb./mL 62 7 205
Bacteria
Total bacteria, nb./mL 2.9×105 1.3×105 5.8×105
Bacteria HNA, nb./mL2 1.9×105 6.6×104 4.2×105
Bacteria LNA, nb./mL2 9.8×104 6.9×104 1.7×105
Chlorophyll
Chlorophyll-a 0.7µm3, µg/L 0.0269 0.0041 0.0388
Chlorophyll-b 0.7µm3, µg/L 0.0030 0.0008 0.0047
Chlorophyll-c 0.7µm3, µg/L 0.0032 0.0007 0.0043
Pheophytin-a 0.7 µm3, µg/L 0.0805 0.0055
SDI SDI3 min 8.0–14.6; Avg. = 11.8 7.1 18.3
SDI5 min 6.5–12.2; Avg. = 9.0 5.4 13.2
SDI15 min 3.7–5.6; Avg. = 4.4 2.6 5.8

1
2/3 of well water + 1/3 of seawater from open intake
2
Bacteria HNA: bacteria with a high DNA fluorescence (bacteria with an elevated cell activity)
Bacteria LNA: bacteria with a low DNA fluorescence (inactive bacteria).
3
1 L sample filtered through a GFF filter with a pore size of 0.7 µm
 J. Leparc et al. / Desalination 203 (2007) 243–255 249

The temperature and conductivity levels of the than that of the well seawater (0.8 mg/L for the
well water and the seawater from the open intake surface seawater as compared to 0.6 mg/L for the
appeared fairly similar, whereas the pH of the well well seawater). These low TOC values are typical
water was slightly lower than that of the surface of the seawater whatever the region. Concerning
seawater. As expected, the turbidity of the well the algal and bacteria content of the raw seawaters,
water was lower than that of surface seawater, and the analyses for picophytoplankton, chlorophyll
the higher particulate content of the seawater from and bacteria show that the rock-well seawater is
the open intake was confirmed with the SDI mea- of a much better quality than the surface seawater.
surements. Indeed, the SDI15 min of the well water The algae and microorganisms concentrations of
was well below 3 whereas the same measurements the rock-well seawater are indeed much lower than
on the seawater from the open intake led to values that of the surface seawater.
around 6. Based on the values of SDI3min and
SDI5min, this difference of water quality between
4. Pretreatment performance evaluation
the well water and the surface seawater was con-
firmed. Fig. 6 compares the SDI3min values of dif-
4.1. SDI abatement
ferent waters. This graph not only demonstrates
the high quality of the well water in Gibraltar but SDI measurements were performed on samples
also shows that the surface seawater in Gibraltar collected at the outlet of the dual-media filters.
is of good quality as compared to the other surface The SDI15-min values averaged 2.1 and ranged
seawaters (Arabian Gulf, Indian Ocean, etc.). between 1.6 and 2.5. Fig. 7 depicts the average
With respect to the organic content of the raw SDI15-min values along the process treatment lane.
seawaters, the TOC and nitrogen concentrations The SDI abatement through the dual-media filters
are low (TOC < 1 mg/L and nitrogen < 0.5 mg/L). is of about 2. The cartridge filters do not improve
The TOC of the surface seawater is slightly higher the SDI.

35

30

25

20
SDI-3 min

15

10

0
Arabian Gulf - Indian Ocean - Mediteranean Gibraltar - well Gibraltar - open Gibraltar -
open intake open intake Sea - open water intake combined raw
intake water

Fig. 6. SDI3min values of different seawater sources.

250 J. Leparc et al. / Desalination 203 (2007) 243–255

6
Average SDI-15min
5

0
Well Surface Combined DMF CF Effluent RO
Seawater Seawater Raw Water Effluent Permeate

Fig. 7. SDI15-min profile along the process treatment lane.

In order to evaluate the ripening of the filters were fairly similar, thus demonstrating that the
with respect to the SDI, samples were collected chlorophyll contribution of the algae passing
during a whole filtration cycle (about 24 h). As through a 3 µm filter but retained on a 0.7 µm fil-
shown by Fig. 8, the SDI15-min value of the filtered ter is negligible. It is also expected that the small
water does not seem to vary significantly during algae passing through the 0.7 µm filter contribute
the filtration cycle. It is interesting to notice that very little to the overall content in chlorophyll
the SDI15-min values of the filtered recorded on and pheophytin.
Filter Cycle 1 are slightly higher that those ob- Pheophytin-a is a degradation by-product of
tained during Filter Cycle 2. This difference is chlorophyll-a. The ratio pheophytin to chlorophyll
explained by the higher SDI values of the raw for all the sample locations ranges between 2 and
seawater recorded during Filter Cycle 1, thus con- 4, thus proving that the chlorophyll-a is partially
firming the impact of raw seawater quality on degraded into pheophytin even before any
performance of direct dual-media filtration. The treatment process (degradation process occurring
SDI15-min values obtained at the outlet of the dual- in the sea). The concentrations of chlorophyll-b
media filters during these two filtration cycles and chlorophyll-c on both filters (0.7 and 3 µm)
ranged between 1.5 and 2.5. were negligible even in the raw waters.
All pigment concentrations are reduced below
4.2. Removal of algae and microorganisms 0.02 µg/l after the dual-media filters, which cor-
respond to a removal rate between 75 and 95%.
4.2.1. Pigments: chlorophyll and pheophytin
Therefore, chlorophyll and pheophytin are effici-
In addition to the standard measurements of ently removed through a single-stage dual-media
chlorophyll retained onto 0.7 µm filters, analyses filtration.
were performed on 3 µm filters in order to evaluate
the chlorophyll contribution of algae of different
4.2.2. Picophytoplankton
sizes. Fig. 9 compares these results along the
treatment lane. The concentrations recorded in the Fig. 10 depicts the concentration profile of the
raw seawaters on the different pore size filters different picophytoplankton species along the
 J. Leparc et al. / Desalination 203 (2007) 243–255 251

2,5
SDI-15 min after DMF

1,5

1
Filter Cycle 1 ; raw water SDI = 12,7 - 13,7
0,5
Filter Cycle 2 ; raw water SDI = 8.0 - 9,1

0
0 5 10 15 20 25
Filtration Time (hour)

Fig. 8. SDI15-min profile of the filtered water during the two filtration cycles.

0,14
Chlorophyll-a 0.7 µm
0,12
Pigment Concentration (µg/L)

Chlorophyll-a 3 µm

0,10 Pheophytine-a 0.7 µm

Pheophytine-a 3 µm
0,08

0,06

0,04

0,02

0,00
Well Surface Combined DMF CF Effluent RO Reject RO
Seawater Seawater Raw Water Effluent Permeate

Fig. 9. Profile and comparison of chlorophyll-a and pheophytin-a concentrations along the treatment lane.

treatment lane. As noted earlier, this graph demon- 28% of Synechococcus (overall picophyto-
strates the much better quality of the rock-well plankton removal of 53%);
seawater as compared to the seawater from the • the 5 µm cartridge filters do not provide further
open intake. The picophytoplankton in the raw removal of the picophytoplankton species;
seawaters is mainly composed of Synechococcus • all the picophytoplankton species are removed
and picoeucaryotes. The concentration profiles through the RO process and the factor of con-
also show that: centration in the reject seems to be respected
• the dual-media filter removed 94% of nano- (factor of concentration of 1.6 related to the
eucaryotes, 87% of picoeucaryotes, and only conversion rate of 39%). Based on these first
252 J. Leparc et al. / Desalination 203 (2007) 243–255

4000
Picophytoplankton Concentration (nb/mL)

3500 Total picophytoplankton (nb/mL)

Synechococcus (nb/mL)
3000
Picoeucaryotes (nb/mL)
2500
Nanoeucaryotes (nb/mL)

2000

1500

1000

500

0
Well Surface Combined DMF CF Effluent RO Reject RO
Seawater Seawater Raw Water Effluent Permeate

Fig. 10. Profile of picophytoplankton species along the treatment lane.

analyses the picophytoplankton does not seem onto the RO membranes. This corresponds to
to be retained in large amount onto the RO a bacteria retention rate on the RO membranes
membranes. Additional analyses should be per- of about 2×109 bacteria/m2 of RO membrane/h.
formed to confirm that observation. Additional analyses should be performed to
confirm that significant retention of bacteria
4.2.3. Bacteria onto the membranes. If the environmental con-
ditions into the RO module promote bacterial
Fig. 11 depicts the concentration profile of
growth, this bacteria retention could lead to
bacteria along the treatment lane. Similarly to the
biofouling.
pigment and picophytoplankton concentrations,
the bacteria enumerations confirm the better
quality of the rock-well water as compared to the
4.3. Removal of the organic matter
surface water. The concentration of bacteria with
high nucleic acid content (HNA) seems to be Fig. 12 depicts the organic carbon detection
higher than the concentration of bacteria with low (OCD)-chromatograms obtained on various sam-
nucleic acid content (LNA). The concentration ples collected along the treatment lane:
profiles also show that: • GOW-RW: Gorge Well raw water (rock-well
• the overall removal rate of bacteria through seawater)
the dual-media filters is lower than 30%, • CP-RW: Camp Bay raw water (surface sea-
• the cartridge filters does not significantly re- water)
move bacteria • Comb-RW: Combined Raw Seawater (raw sea-
• the bacteria concentration in the brine is slight- water after blending and entering the plant)
ly lower that the one expected based on the • DMF: effluent water from dual-media filters
concentration factor. It seems that 30–40% of • CF: effluent water from cartridge filters
bacteria entering the RO system are retained
 J. Leparc et al. / Desalination 203 (2007) 243–255 253

Total Bacteria
1000000 Bacteria HNA
Bacteria LNA

100000
Bacteria Concentration (nb/mL)

10000

1000

100

10

1
Well Surface Combined DMF CF Effluent RO Reject RO
Seawater Seawater Raw Water Effluent Permeate

Fig. 11. Profile of bacteria concentrations along the treatment lane (log-scale).

1,9 Well Seawater - 0,82 ppmC

GOW-RW-1 Surface Seawater
CP-RW-1- 0,92 ppmC
Combined
Comb-RW-1 Raw-Seawater
0,78 ppmC Dual Media
DMF-2 Filtration
- 0,89 ppmC Effluent
Cartridge Filtration
CF-1 - 0,92 ppmCEffluent

1,4

0,9
OCD

0,4

-0,1

-0,6
0 8 17 25 33 42 50 58 67 75 83 92 100 108 117 125 133
time (min)

Fig. 12. LC-OCD chromatograms from samples collected along the treatment lane.

All the chromatograms seem to have the same nificantly the TOC content. The UV254 value was
profile, and the resulting TOC values (data shown reduced only from 0.68 m–1 to 0.5 m–1 through
on the right-hand side of the legends) are similar the dual-media filters. The cartridge filters did not
as well. These results show that the dual-media provide additional abatement of UV254.
filters and the cartridge filters do not reduce sig- To differentiate the fractions of the natural
254 J. Leparc et al. / Desalination 203 (2007) 243–255

organic matter, the following analysis and inter- content. Regarding the difference in organic mat-
pretation of the chromatograms can be made: ter composition of the rock-well seawater and the
• the first fraction detectable in OC detection be- surface seawater in Gibraltar, the content in humic
tween 30 and 50 min (first peak of the chroma- substances and lower molecular weight acids and
togram, largest molecular size) is the polysac- neutrals are similar. Low molecular weight com-
charide and protein peak, pounds are present in a greater proportion in the
• the second fraction between 50 and 70 min is rock-well seawater, whereas the surface seawater
attributed to organic molecules of high molec- contains more polysaccharides as compared to the
ular weight like humic substances and the cor- rock-well seawater. This demonstrates that partial
responding building blocks. biodegradation of the high molecular weight com-
• the third peak (between 70 and 77 min) cor- pounds (i.e. polysaccharides) may occur in the
responds to the low molecular mass acids and process of natural slow filtration through the sea-
neutrals, and bed. Regarding the impact of the pretreatment
• the last fraction (between 77 and 110 min) cor- process on the natural organic matter, Figs. 12 and
responds to the smallest molecular weight 13 confirm that the dual media filtration and the
compounds. cartridge filtration at the Gibraltar SWRO plant
do not impact significantly the composition of the
Fig. 13 compares the different organic fractions organic matter.
of the samples collected in Gibraltar with those
of raw seawaters from other sites (Indian Ocean,
5. Conclusions
Mediterranean Sea). The low molecular weight
compounds represent the greatest fraction (be- Advanced water quality analyses were used
tween 40 and 50%) of the overall organic carbon to characterize the raw seawater feeding a SWRO

Polysaccharides
60 Humic Substances and Building Blocks
Low Molar Mass Acids + Neutrals
Low Molar Mass Compounds
50

40
Fraction %

30

20

10

0
Surface Raw Surface Raw Well Seawater Surface Raw Combined DMF Effluent - CF Effluent -
Seawater - Seawater - - Gibraltar Seawater - Raw Seawater Gibraltar Gibraltar
Indian Ocean Mediteranean Gibraltar - Gibraltar
Sea

Fig. 13. Comparison of the organic fractions from samples of various origins.
 J. Leparc et al. / Desalination 203 (2007) 243–255 255

plant and also allow a better assessment of the 6. No significant removal of bacteria through dual
pretreatment performance. The two raw seawater media and cartridge filtration was recorded.
sources feeding the Gibraltar SWRO plant showed 7. No organic removal and no change of the or-
significant differences of water quality. Indeed, ganic matter characteristics through dual media
the seawater from the rock-well had a much better and cartridge filtration were observed.
water quality based on a wide variety of parame-
ters (turbidity, SDI, bacteria, picophytoplankton, Overall, the pretreatment process appears to
chlorophyll). be effective to provide high water quality to the
Regarding the efficiency of the pretreatment, reverse osmosis units. The high quality of the pre-
the following observations can be made: treated water is not only demonstrated by the
1. The pretreatment process composed of a results of the water analyses but also by the fact
single-stage dual-media filtration provides a that no major fouling events have occurred since
SDI abatement from 4.4 to 2.1. The cartridge the beginning of operation (September 2003).
filters do not provide additional SDI abate-
ment.
2. No significant variation of SDI in the filtered References
water was observed during a filtration cycle.
[1] A.J. Karabelas, Critical Assessment of Fouling
3. The SDI of the filtered water seems to mainly Indices. MEDRC Report, Project #98-BS-034, 2003.
depend on the feedwater quality: increasing [2] S.F.E. Boerlage, M. Kennedy, Z. Tarawnwh, R. De
feedwater SDIs lead to higher filtered water Faber and J.C. Schippers, Development of the MFI-
SDIs. UF in constant flux filtration. Desalination, 161
4. Algal pigments (chlorophyll and pheophytin) (2003) 103–113.
are efficiently removed through dual media fil- [3] Plant Performance Normalization. Filmtec Mem-
tration. branes Technical Manual.
[4] Standard Method for Silt Density Index of Water,
5. The removal of picophytoplankton depends on
ASTM D 4189-95.
the species.