ARTICLE IN PRESS
Water Research 39 (2005) 232–238
www.elsevier.com/locate/watres
The oxygen isotope composition of dissolved anthropogenic
phosphates: a new tool for eutrophication research?
Gérard Gruaua,, Michèle Legeasb, Christine Rioua, Eve Gallacierb, Francois Martineauc, O. Hénina
a
Geosciences Rennes, CAREN, CNRS UMR 6118, Campus de Beaulieu, 35042 Rennes Cedex, France
b
ENSP, Avenue Léon Bourgeois, 35042 Rennes, France
c
Université de Lyon 1, CNRS UMR 5125, 69622 Villeurbanne Cedex, France
Received 7 April 2004; received in revised form 26 August 2004; accepted 30 August 2004
Abstract
High-precision oxygen isotope analyses were carried out on dissolved phosphate extracted from discharge waters
from three wastewater treatment plants (WTP) located in western France, as well as on the different phosphate-based
fertilizers applied by farmers in the same region. Measured d18O values of phosphate from chemical fertilizers range
from 19.6 to 23.1%, while those of phosphate from WTP discharge waters are more tightly grouped between 17.7 and
18.1%. The variablility in d18O values of phosphate fertilizers is attributed to oxygen isotope variations of the
phosphorite deposits from which France’s fertilizers are manufactured. The significance of the d18O values of phosphate
from WTP discharge waters is less straightforward. At present, it is not clear whether these values are primary isotopic
compositions corresponding, e.g., to the oxygen isotope composition of phosphate builders included in detergents
(d18OP=17.9%), or represent secondary values reflecting biological recycling of the phosphate in equilibrium with
ambient WTP water The restricted difference in isotopic composition obtained between phosphate from fertilizers and
phosphate from WTP discharge waters (o2%), as well as the fairly large internal isotopic variability observed in both
end-members (X1.5%), cast doubt about the possibility that the oxygen isotope composition could serve as a tracer for
the source of anthropogenic phosphates in waters.
r 2004 Elsevier Ltd. All rights reserved.
Keywords: Phosphate; Oxygen isotope; Eutrophication; Sewage; Fertilizer
1. Introduction
It is well established that anthropogenic increase in
the amount of phosphate in surface waters can lead to
eutrophication and can damage the water quality. Since
the 1950s, increased application of phosphate-based
Corresponding author. Tel.: +33 2 23 23 60 86;
+33 2 23 23 14 99.
E-mail address: [email protected] (G. Gruau).
fax:
fertilizers in agriculture (including manure) has given
rise to substantial phosphate additions in lakes and
rivers, the so-called diffuse or non-point source pollution (Novtony, 1999; and references therein). Meanwhile, more and more nutrients have been channelled
through the food chain into the sewage systems, and
have been discharged into surface waters through pipes,
the so-called point source pollution (Forsberg , 1998). In
some cases, predominance of sewage in the eutrophication process is obvious from the coupling between
0043-1354/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2004.08.035
ARTICLE IN PRESS
G. Gruau et al. / Water Research 39 (2005) 232–238
2. Experimental methods
The first method of isolating dissolved phosphate was
published by Longinelli et al. (1976). This is a three-step
method which uses natural sponges impregnated with
iron hydroxide to quantitatively remove phosphate ions
from sample solutions. The phosphate ions are then
leached by nitric acid, purified, and finally precipitated
as BiPO4. This is a relatively laborious method which
suffers from the great disadvantage of ending up by
precipitating a hygroscopic form of BiPO4 crystals. The
method presented here combines the uptake of phos-
phate by iron hydroxide precipitates and the Ag3PO4
technique developed by Lécuyer et al (1993).
2.1. Precipitation of dissolved phosphate with iron
hydroxide and purification as Ag3PO4
Between 500 and 300 ml of WTP water is successively
filtered through membranes of 2 mm and 0.45 mm pore
size. Samples are then passed through a 200 ml column
of activated C to remove dissolved organic carbon
(DOC). DOC is removed because it can contain up to
20 wt% O that could interfere with the phosphate
isotope analysis. The DOC-free samples are then placed
in a 1 l beaker and a phosphate–iron hydroxide
precipitate is formed at room temperature by the
addition of 0.1 M FeSO4. During this addition, the pH
of the solution generally falls to a value of about 6.5. As
shown in Fig. 1, the uptake of phosphate by iron ferrous
hydroxides is maximum (495% yield) when the pH of
the solution is set to a value of 8.570.1. Thus, a few ml
of NaOH is systematically added to raise the pH of the
sample. The solution is then gently shaken for about
20 min and the greyish-green precipitate allowed to settle
in the beaker over a period of 24 h. Finally, the
surpernatant is gently removed by slow aspiration and
the precipitate placed in an oven and dried overnight at
30 1C.
After drying, 10 ml of 2 M KOH is added and the
sample transferred to a polypropylene tube. The tube is
placed on a shaker table for 5 h to promote desorption
of the phosphate ions. The iron hydroxide is separated
from the phosphate solution by centrifugation. The iron
hydroxide precipitate is rinsed two times with 5 ml of
KOH, and the rinsed KOH solution added to the
phosphate solution. The phosphate is then purified using
cleaned Amberlite-IRA-400s (OH form) ion exchange
resin, and finally isolated as Ag3PO4 crystals.
Uptake (%Psol)
urbanization/industrialization of a catchment and the
subsequent eutrophication of receiving waters. In most
cases, however, it is unclear whether sewage or
agriculture contributes most intensely to the eutrophication process (Lung, 1996).
Recent studies showed that the oxygen isotope
composition of nitrates could help in distinguishing
nitrate sources in ground and surface waters, particularly in cases where nitrate-based fertilizers are used
(Durka et al., 1994; Wassenaar, 1995). As for nitrates,
phosphorus is added to waters as oxygen-bound
phosphorus (i.e. phosphate), thus allowing the oxygen
isotope content to perhaps be also used as a tracer of
phosphate sources in waters. This potential was first
tested by Markel et al. (1994). The studied site was Lake
Kinneret in Israel, the analyzed samples being phosphate from both the sediments and the suspended matter
of this lake. Using the oxygen isotope tool, Markel and
co-workers (1994) were able to estimate that ca. 60% of
the phosphate that enters Lake Kinneret comes from an
anthropogenic source, even though they failed to
establish the exact nature of this source.
A prerequisite condition of all isotopic work is that
measurable isotopic differences must exist among the
various sources of the element or molecule that is to be
traced. In the present case, this means that one must first
establish that sewage and agricultural phosphates have
distinct oxygen isotope signatures. Moreover, a reliable
analytical method, suitable to analyze the oxygen
isotope composition of dissolved phosphate (i.e., the
phosphate form which is directly bioavailable to
phytoplanktonic communities) needs to be developed.
In this context, the specific objectives of the present
work are: (i) to set up a method suitable to determine the
oxygen isotope composition of dissolved phosphate, (ii)
to use this method to measure the oxygen isotope
composition of phosphates from urban wastewater
treatement plants (WTP) located in an agricultural
region from western France, and (iii) to compare the
WTP phosphate data with results obtained on the
phosphate-based fertilizers (superphostate, NP, PKS)
that are used by the farmers from the same region.
233
100
90
80
70
60
50
40
30
20
10
0
5
6
7
8
9
10
11
12
pH
Fig. 1. pH dependence of phosphate uptake by ferrous
hydroxide at 20 1C. Error bars give estimates of measurement
precision.
ARTICLE IN PRESS
G. Gruau et al. / Water Research 39 (2005) 232–238
234
2.2. Fluorination decomposition of Ag3PO4 and mass
spectrometry
24
2.3. Sample size, blank effects, yield and reproducibility
of the method
Phosphate yields of the whole method and reproducibility of d18O measurements were checked by means of
a synthetic phosphate solution prepared using a
phosphate fertilizer sample of known oxygen isotopic
composition (d18O value=21.970.2%). As can be seen
in Table 1, the yield of the overall method is typically
80–90% with an average reproducibility of about
70.2%. Moreover, the average measured value is
22.170.4% (1s, n ¼ 10), which is within error of the
isotopic composition of the material used to prepare the
synthetic phosphate solution.
Another series of tests was made in order to establish
the minimum amount of Ag3PO4 crystals that should be
Table 1
Yield, reproducibility and accuracy of the iron hydroxide—
Ag3PO4 method
Sample
Yield (%)
d O values (%)
1
2
3
4
5
93
81
87
79
90
22.2
22.6
22.4
22.2
22.2
a
21.8
21.2
22.1
21.8
22.0
Ag3PO4 crystals were obtained using a synthetic phosphate
solution prepared by dissolving 30 mg of NP fertilizer
(d18O=21.970.2%; see Table 2) in 1 l of distilled water.
a
Values obtained using two different aliquots of the same
sample. Average error is 70.25%.
δ18OPO4 (‰)
Weighted aliquots between 10 and 15 mg of Ag3PO4
crystals are loaded into nickel reaction vessels and
degassed for 2 h at room temperature and an additional
3 h at 150–200 1C to desorb all traces of atmospheric
water. The silver phosphate is then reacted with a 5/
1 mole excess of BrF5 at 650 1C for 12 h, the extracted O2
being converted to CO2 by reaction with a heated
graphite rod. The oxygen and carbon isotopic composition of the CO2 gas is finally measured using a VG
SIRA-10s mass spectrometer and the results quoted in
the standard d notation relative to the VSMOW
reference value.
Replicate analyses of 19 samples of the NBS 120c
Florida Standard yield a mean d18O value of +21.9%
(1s ¼ 0:3) with a standard deviation of 70.3%. The
NBS 28 quartz that has been extensively analysed in
several laboratories was also used as a control on the
fluorination procedure and mass spectrometry runs. The
mean d18O value obtained for this standard is
9.370.1%.
18
21.9± 0.2‰
22
20
Error 1
18
16
Blank
14
NBS 120c
12
-2
0
2
4
6
8
10 12
Weight of Ag3PO4 (mg)
14
16
Fig. 2. Plot showing minimum size of Ag3PO4 sample that
must be prepared and analyzed to overcome background
contamination.
loaded in the silicate extraction lines at Rennes in order
to overcome the background contamination. Indeed,
this minimum amount, if large, could become a major
complication if the method presented above was to be
applied to natural waters (e.g., rivers, lakes, runoff, etc.),
as orthophosphate concentrations in these waters are
often very low, typically o0.1 ppm (Lung, 1996;
Sharpley et al., 1995). To establish this minimum sample
size, 20 aliquots between 1 and 15 mg of Ag3PO4 crystals
were prepared from the NBS 120c Florida standard
(mean isotopic composition of 21.970.3%) and analyzed for their 18O/16O ratio. The blank extracted from
the silicate extraction lines used at Rennes has a d18O
value of 13% (Fig. 2), so that if blank contamination
took place during fluorination, the samples prepared
from the NBS 120c Florida standard would have their
isotopic composition shifted toward valueso21.9%. As
can be seen in Fig. 2, this arises when the crystal weight
is about 4 mg, indicating that 4 mg corresponds to the
minimum amount of Ag3PO4 crystals that can be
confidently and reliably run.
3. Samples
The site chosen to test the ability of the oxygen
isotope tool to trace anthropogenic phosphate sources is
composed of three small, contiguous agricultural districts—Vézin le Coquet, La Mézière and Pacé—located
in Brittany, approximately 350 km west of Paris, France.
The entire Brittany region is one of the most productive
French and European agricultural regions accounting
for ca. 40% of the French pig production. The problem
of eutrophication of surface waters became apparent in
Brittany about two decades ago when the amounts of
phosphate and nitrate released by agriculture and urban
wastewater discharges (domestic and industrial) started
to increase dramatically. In 1987, 50% of the water
ARTICLE IN PRESS
G. Gruau et al. / Water Research 39 (2005) 232–238
resources used in Brittany for water supply were affected
by eutrophication, a situation which has not evolved
favoarably since, due to increased intensification of
agricultural practices (Soulard, 1994). Today, the entire
Brittany area is mapped as a sensitive zone with regard
to eutrophication according to the European Union
directive on sewage. In Brittany, the main sources of
phosphorus added to agricultural lands are artificial
phosphate-based fertilizers (tricalcic phosphate, diammonium phosphate, etc.) followed by animal manure,
with a mean annual input of phosphorus from artificial
fertilizers averaging 4 tons of P/km2/year (Soulard,
1994).
Three different types of artificial fertilizers sourced
from the two main agricultural co-operatives, which
deliver artificial fertilizers in the studied districts were
Table 2
Oxygen isotope composition of anthropogenic phosphates
Phosphate
content
Chemical fertilizers (bulk
analyses)
1/NP/Coopagri
2/NP/Coopagri
3/NP/Coopagri
1/NP/Coralis
2/NP/Coralis
1/SuperP/Coopagri
2/SuperP/Coopagri
1/PKS/Coralis
2/PKS/Coralis
(%PO4)
33.5
33.5
33.5
33.4
33.4
30.5
30.5
7.3
7.3
Chemical fertilizers (dissolution experiments)
Fraction 1
30c
Fraction 2
50c
Fraction 3
100c
WTP Discharge Waters
Plant location
La Mézière
Pacé
Vézin le Coquet
Vézin le Coquet
Vézin le Coquet
Phosphate Builder from
Detergents
(mg PO4/
liter)
33
28
25
33
27
n.d.
analyzed (Table 2). As most other agricultural regions of
western Europe, all of France’s artificial fertilizers are
manufactured from imported phosphorites. In France,
the dominant sources of imported phosphorites are
Morocco, USA (Florida), Tunisia and Senegal in that
order in volume.
The three WTP from which dissolved phosphate was
analyzed for purpose of comparison have the same
overall treatment capacities: they seave on 5000 to 10000
equivalent inhabitants each, and all these plants use the
same treatment system of the activated sludge technique.
Treated sewages are essentially of domestic origin and
none of the studied plants performs iron treatments of
its effluents. Samples were collected at the outlet of each
plant. Measured dissolved phosphate concentrations
were fairly similar, ranging from 20 to 33 mg PO4.l1
(Table 2).
4. Results and discussion
18
d OPO4
22.170.2a
22.070.1a
21.970.1a
21.470.1b
21.770.1b
23.170.2a
22.970.1a
19.670.1b
19.770.1b
4.1. Phosphate from fertilizers
Phosphate oxygen in artificial fertilizers was analyzed
following the Ag3PO4 method developed by Lécuyer et
al. (1993) for solid materials. As shown in Table 2 and
Fig. 3, the analyzed samples gave different oxygen
isotope compositions, ranging from 19.6 to 23.1%. The
d18O values measured in these fertilizers are within the
range of the values found in upper Cretaceous and
Tertiary phosphorites from Morocco and Florida
12
22.370.1b
21.570.1b
21.670.1b
18.470.1b
16.670.1b
17.770.1b
18.070.1b
17.670.1b
17.970.1b
Total uncertainties with respect to the VSMOW scale are
estimated to be lower than 70.25.
n.d.: not determined
a
Errors corresponding to the reproducibility of three separate
extractions and mass spectrometry analyses.
b
Errors obtained during mass spectrometry analyse.
c
These numbers are not concentrations, but indicate relative
mass fraction of dissolved phosphate.
Number of analyses
Sample
235
10
8
6
Morocco
18.5‰
20.5‰
Florida
17.2‰
23.2‰
4
2
0
16
17
18
19
δ18O
20
21
22
23
24
PO4 (‰)
Fig. 3. Plot showing a comparison between the d18O values of
the phosphate fertilizers investigated in this study (squares) and
the range of values published for sedimentary phosphate
deposits from Morocco and Florida (lines). The figure shows
that both sets of values overlap each other, thus confirming the
suggestion made earlier (Durka et al., 1994) that the oxygen
isotope composition of phosphate in artificial fertilizers is
controlled by that of the sedimentary phosphate deposits from
which they are manufactured.
ARTICLE IN PRESS
236
G. Gruau et al. / Water Research 39 (2005) 232–238
(Lécuyer et al., 1993; Longinelli and Nuti, 1968; Karhu
and Epstein, 1986), thus confirming the suggestion made
earlier (Durka et al., 1994) that the oxygen isotope
composition of phosphate in artificial fertilizers is
controlled by sedimentary phosphate deposits from
which the fertilizers are manufactured. The results
presented here also suggest that the different manufacturing processes used to prepare secondary and ternary
phosphate fertilizers have essentially no effect on the
original oxygen isotope compositions.
A progressive dissolution experiment was carried out
to test whether dissolution of the phosphate fertilizers
during, e.g., a rainfall event could change their original
oxygen isotope composition. In this experiment, about
100 mg of a NP fertilizer (d18O=21.770.3%) was
placed in a 100-ml reservoir, and dissolved progressively
using distilled water (d18O=6%) at 20 1C. Samples of
the eluted phosphate solutions were recovered and
analyzed for their oxygen isotope composition at t ¼
10; 20 and 40 min. As can be seen in Table 2, measured
d18O values of all three leachates were identical, within
analytical error, to that of the solid phosphate-based
fertilizer used in the dissolution experiment.
controlled values, but represent secondary isotope
compositions arising from the biological cycling and
biological re-equilibration of the WPT phosphate by the
bacteria involved in the sewage treatment processes.
Blake and co-workers (1997) reported results of
laboratory experiments on enzyme-mediated reaction
of phosphate and microbially mediated degradation of
organic matter, a mechanism which represents the basis
for the sewage treatment processes used in the three
studied WTPs. Their results demonstrate that significant
exchange of oxygen isotopes between phosphate and
water accompanies the hydrolitic cleavage of organically
bound phosphate as well as the metabolism of inorganic
phosphate, and this exchange can result in complete reequilibration of oxygen isotopes between phosphate and
water. In Fig. 4, we present calculated curves, which
show what the d18O values of our WTP phosphates
should be, if the microbially mediated phosphate reequilibration took place. The equation used is the
microbially mediated fractionation equation given by
Blake et al. (1997)—[T(1C)=155.8 6.4 (103ln a)], where
a represents the fractionation factor between phosphate
and water. Three modelled waters with d18O values
4.2. Phosphate from WTP
20
- 4‰
Vézin
19
δ18OPO4 (‰)
WTP phosphate samples have d18O values significantly lower and more tightly grouped than those
measured in fertilizers (Table 2). Interpretation of the
WTP results is not straightforward as many different
sources of phosphate may ultimately contribute to the
total phosphate pool that is discharged from these
plants. Mass balance calculations suggest that human
wastes such as feces, urine and waste food disposal could
account for about 30–50% of the phosphate budget, the
remaining 50–70% consisting of phosphate released
from phosphate-based detergents. Thus, it is expected
that the phosphate pool discharged from these WTPs
will consist of a complex mixture between different,
initially organically bound phosphate components and
an inorganic phosphate component derived mainly from
domestic detergents. At present, data are lacking to
ascribe reliable d18O values to each of these components
and it is consequently difficult to evaluate whether the
WTP d18O values reported here are source-controlled
values. The only constrain we have comes from the
analysis of a phosphate builder that has a d18OP value of
17.9% (Table 2). This value is clearly within the range of
the d18O measured in WTP phosphate samples (from
16.6 to18.1%). Although this value of 17.9% might not
be representative of the oxygen isotope composition of
all the phosphate builders that enter the three studied
WTPs, it does support the possibility that the WTP d18O
values reported here could be source-controlled values.
However, another possibility is worth considering,
namely, that these values are not primary, source-
Pacé
- 5‰
La Mézère
18
- 6‰
17
16
15
0
5
10
15
Température ˚C
20
25
Fig. 4. Comparison between the d18O values measured in WTP
phosphates and the values expected assuming complete,
biologically mediated, isotopic equilibrium between the oxygen
in the phosphate and the oxygen in the WTP water. Three
waters were tested (d18O=4, 5, and –6%), corresponding to
the range of d18O values measured in local surface waters. The
three straight lines showing the oxygen isotopic compositions of
dissolved, biologically equilibrated phosphate as a function of
equilibrium temperature were calculated using the fractionation
equation given by Blake et al. (1997). The points refer to the
results obtained from the La Mézière, Pacé and Vézin WTPs
(i.e. oxygen isotope composition of dissolved phosphate at the
WTP outlets and temperature measured in the WPT water at
time of phosphate sampling). The fact that the La Mézière,
Pacé and Vézin WTP points overlap the ‘‘biological’’ theoretical
phosphate lines makes it possible that the WTP values
measured in this study could not be primary source values
(e.g. values of the phosphate builders included in detergents),
but could represent secondary values reflecting rapid biological
recycling of phosphate during water treatment process.
ARTICLE IN PRESS
G. Gruau et al. / Water Research 39 (2005) 232–238
encompassing the range of local surface waters d18O
(from –6 to –4%; (Mérot, 1995)) were used for the
calculation. It is clear from Fig. 4 that d18O values in the
range 16.6 to 18.1% values obtained here correspond
closely to the d18O that are calculated in case of a
complete, microbially mediated, isotopic re-equilibration between phosphate and ambient WTP waters, for
temperatures in the range 10–15 1C (i.e. the actual
temperature range measured during sampling).
At present, we cannot decide whether the d18O values
recorded in the La Mézière, Vézin le Coquet and Pacé
WTPs should be regarded as source-controlled values
or, instead, as secondary, re-equilibrated values. A
determination of the exact origin and exact significance
of the WTP values would require careful separation and
analysis of the different phosphate sources that enter
these WTPs, as well as more experimental constraints on
the isotopic cycling of phosphate in WTP.
4.3. Can the oxygen isotope composition of dissolved
phosphate be used as a tracer of phosphate sources?
We now turn to the starting question: can the oxygen
isotope composition of phosphate be a suitable tracer of
the source of anthropogenic phosphates in aquatic
ecosystems? It should be realized immediately that the
isotopic difference obtained here between WTP phosphates and fertilizer phosphates (d18O from 16.6 to
18.1%, and from 19.6 to 23.1%, respectively), though
statistically real, is quite small being o2d units. Moreover, it is to be noted that the internal variability of
oxygen isotope compositions is quite large in both endmembers (X1.5 d units in both cases; see Table 2). This
variability in the end-member fingerprinting coupled
with the relatively small difference in d18O, if confirmed
elsewhere, would severely limit the use of the oxygen
isotope compositions as a tool for distinguishing exact
amounts as it would result in large uncertainties in the
calculated proportions. This is illustrated in Table 3,
where we have simulated the case of an eutrophic lake
that would be contaminated by a phosphate fertilizer
comparable to that analyzed in this study and a WTP
phosphate with the same range in isotopic composition
as that measured in the Pacé, Vézin le Coquet and La
Mézière WTPs. In Table 3, the d18O value of the lake
phosphate is assumed to be constant and equal to 19 per
mil. Table 3 shows that fertilizer proportions as variable
as 20 to 80% may be calculated, depending on which
d18O values are ascribed to the WTP and fertilizer endmembers. As mentioned in introduction, one of the most
recurrent questions in eutrophication control is to know
how much phosphate in the water column at a certain
location is from an agricultural source and how much
comes from an urban or industrial source. It is not
entirely certain that an uncertainty as big as that
reported in Table 3 on phosphate provenance will be
237
Table 3
Calculated proportions of fertilizer and WTP sourced phosphate
d18OWTP(%)a
d18OFertilizer (%)a
Calculated proportion
of phosphate fertilizer
in lakeb
+16.6
+18.1
+16.6
+18.1
+23.1
+23.1
+19.6
+19.6
38%
20%
80%
60%
a
Values used in the mixing equation for the sewage
(d OWTP) and fertilizer (d18OFertilizer) phosphate end-members.
The d18O in the model eutrophic lake receiving phosphate from
both sources is assumed to be equal to +19%.
b
Values calculated using the following mixing equation:
18
d OLake=d18OFertilizer . P+d18OWTP . (1P), where P is the
proportion of phosphate from fertilizers, d18OPlake is the
isotopic composition of phosphate in the lake, and d18OFertilizer
and d18OWTP are the isotopic compositions of fertilizer and
sewage phosphate end-members, respectively.
18
judged acceptable by the legal entities in charge of
evaluating the impact of current and proposed agricultural practices on the amount of phosphate which is
transferred from cultivated lands to surface water
masses.
5. Conclusion
A reliable analytical method, suitable to analyze the
oxygen isotope composition of dissolved phosphate is
developed. The method, which combines the uptake of
phosphate by iron hydroxide precipitates and the
classical Ag3PO4 technique used in phosphate isotope
studies, allows measurement of the oxygen isotope
composition of ca. 1 mg of dissolved PO4 with a
precision better than 7 0.5%.
This method was used to test the potential of the
oxygen isotope content of phosphate to distinguish
between phosphate from phosphate-based fertilizers and
phosphate from treated domestic sewages. Measured
d18O values of phosphate from chemical fertilizers range
from 19.6 to 23.1%, while those of phosphate from
sewages are more tightly grouped between 17.7 and
18.1%. The fertilizer results confirm earlier suggestions
that the oxygen isotope composition of phosphate-based
fertilizers is controlled by that of the sedimentary
phosphate deposits from which the fertilizers are
manufactured. The d18O values of phosphate from
treated sewages are identical to the oxygen isotope
composition of phosphate builders included in detergents (d18OP=17.9%), and could thus represent sourcecontrolled values. However, these values are also rather
ARTICLE IN PRESS
238
G. Gruau et al. / Water Research 39 (2005) 232–238
close to the values that may be calculated by means of
the microbially mediated, phosphate–water equation of
Blake et al. (1997), raising the possibility that they could
represent secondary values stemming from the biological
recycling of sewage phosphate sources by the bacteria
involved in the sewage treatment processes.
At any rate, the restricted difference in isotopic
composition obtained in this study between phosphate
from fertilizers and phosphate from treated sewage
(o2%) casts doubt about the possibility that the oxygen
isotope composition of phosphate could serve as a tracer
for the source of anthropogenic phosphates in waters.
Acknowledgments
This work was initiated through discussions with C.
Lécuyer and S. Fourcade. C. Lécuyer is gratefully
acknowledged for his assistance during sample preparation and mass spectrometry analyses. This work was
supported by the French Ministry of Environment and
the Regional Council of Brittany.
References
Blake, R.E., O’Neil, J.R., Garcia, G.A., 1997. Oxygen isotope
systematics of biologically mediated reactions of phosphate:
I. Microbial degradation of organophosphorus compounds.
Geochim. Cosmochim. Acta. 61, 4111–4122.
Durka, W., Schulze, E.-D., Gebauer, G., Voerkelius, S., 1994.
Effects of forest decline on uptake and leaching of deposited
nitrate determined by 15N and 18O measurements. Nature
372, 765–767.
Forsberg, C., 1998. Which policies can stop large scale
eutrophication? Water. Sci. Tech. 37, 193–200.
Karhu, J., Epstein, S., 1986. The implication of the oxygen
isotope records in coexisting cherts and phosphates.
Geochim. Cosmochim. Acta. 50, 1745–1756.
Lécuyer, C., Grandjean, P., O’Neil, J.R., Capetta, H.,
Martineau, F., 1993. Thermal excursion in the ocean at
the Cretaceous–Tertiary boundary (northern Morocco):
d18O record of phosphatic fish debris. Palaeogeogr. Palaeoclimatol. Palaeoecol. 105, 235–243.
Longinelli, A., Nuti, S., 1968. Oxygen isotopic composition of
phosphorites from marine formations. Earth Planet. Sci.
Lett. 5, 13–16.
Longinelli, A., Bartelloni, M., Cortecci, G., 1976. The isotopic
cycle of oceanic phosphate: I. Earth Planet. Sci. Lett. 32,
389–392.
Lung, W.S., 1996. Fate and transport modeling using a
numerical tracer. Water Resour Res 32, 171–178.
Markel, D., Kolodny, Y., Luz, B., Nishri, 1994. Phosphorus
cycling and phosphorus sources in lake Kinneret: tracing
by oxygen isotopes in phosphate. Isr. J. Earth. Sci. 43,
165–178.
Mérot, P., Durand, P., Morisson, C., 1995. Four-component
hydrograph separation using isotopic and chemical determinations in agricultural catchments in western France.
Phys. Chem. Earth 20, 415–425.
Novtony, V., 1999. Diffuse pollution from agriculture—a
worldwide outlook. Water. Sci. Tech. 39, 1–13.
Sharpley, A.N., Hedley, M.J., Sibbesen, E., Hillbricht-Ilkowska, A., House, W.A., Ryszkowski, L., 1995. Phosphorus
transfers from terrestrial to aquatic ecosystems. In: Tiessen,
H. (Ed.), Phosphorus in the Global Environment. Wiley,
London, pp. 171-199 (Chapter 11).
Soulard, B., 1994. Phosphore et eutrophisation en Bretagne. In:
Impact du phosphore dans l’environnement et prévention de
l’eutrophisation, Journées d’Echanges Techniques de Derval
(44), DDASS de Loire-Atlantique et d’Ille et Vilaine, pp.
15–20.
Wassenaar, L.I., 1995. Evaluation of the origin and fate of
nitrate in the Abbotsford Aquifer using the isotopes of 15N
and 18O in NO3. Appl. Geochem. 10, 391–405.