Environmental Pollution 132 (2004) 413426

www.elsevier.com/locate/envpol

Fifty-year sedimentary record of heavy metal pollution

(Cd, Zn, Cu, Pb) in the Lot River reservoirs (France)
Stephane Audry*, Jorg Schafer, Gerard Blanc,
Jean-Marie Jouanneau
Traceurs Geochimiques et Mineralogiques (T. G. M.) team, Universite Bordeaux 1, UMR CNRS 5805 EPOC,
Avenue des facultes, 33405 Talence cedex, France
Received 13 February 2004; accepted 21 May 2004

Capsule: Reservoir sediments provide ne-scale information on the historical

record of metal pollution in a watershed.
Abstract
The LotGaronne uvial system is known for its historic heavy metal pollution resulting from mining and smelting activities
since the late 19th century. Here, we report 137Cs activities and heavy metal (Cd, Zn, Cu, Pb and V) concentrationdepth proles
from sediment cores retrieved in 2001 from three reservoirs in the Lot River. High mean sedimentation rates of 2.42.8 cm a
1 are
indicated by 137Cs dating. The reservoir sediments have recorded the heavy metal deposition and thus allow establishing
a connection between the temporal evolution of the heavy metal pollution and historical changes in smelting and waste-treatment
proceedings. Based on heavy metal concentrations in sediments upstream of the anthropogenic inputs and bottom-sediments of the
furthest downstream core (interpreted as old soil or riverbed), concentrations of w17, w82, w0.33 and w28 mg kg
1 for Cu, Zn,
Cd and Pb, respectively, are proposed as natural background values for the Lot uvial system. The geoaccumulation index (Igeo
[Muller, G., 1979. Schwermetalle in den Sedimenten des Rheins-Veranderungen seit. Umschav 79, 133149.]) revealed that the Lot
River sediments must be considered as severely polluted in Cd and Zn. Moreover, despite remediation eorts undertaken in the
former smelting site, the Lot River is still severely (Igeo w 4) and moderately to severely (Igeo O 2) impacted by Cd and Zn
inputs, respectively.
2004 Elsevier Ltd. All rights reserved.
Keywords: Heavy metal; Pollution; Sediment; Record; Cesium

1. Introduction
Anthropogenic activities have caused important
transformations in aquatic environments during the last
150 years. Heavy metals are among the most widespread

* Corresponding author. Tel.: C33 540008834.

E-mail addresses: [email protected] (S. Audry), j.schaefer@
epoc.u-bordeaux1.fr (J. Schafer).
0269-7491/$ - see front matter 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envpol.2004.05.025

of the various pollutants originating from anthropogenic activities, particularly from mining and smelting waste
sites (e.g. Salomons, 1995; Hochella et al., 1999). Such
sites typically release large amounts of heavy metals into
uvial systems that can be transported along hydrologic
gradients for hundreds of kilometers in relatively short
times. The release of heavy metals can occur both in
dissolved and particulate form. When released in
particulate form or adsorbed to particles, heavy metals
can settle down and be deposited in the bottom river

414

S. Audry et al. / Environmental Pollution 132 (2004) 413426

sediments under favorable hydraulic conditions provided by natural and reservoir lakes. Thus, historical
records of the anthropogenic emissions into uvial
environments can be provided by lacustrine or reservoir
sediments (e.g. Von Gunten et al., 1997; Kahkonen
et al., 1998; Muller et al., 2000; Arnason and Fletcher,
2003). Reservoir sediments usually accumulate at high
rates (typically O2 cm a
1; Muller et al., 2000; Arnason
and Fletcher, 2003) compared to lacustrine sediments
(often !0.3 cm a
1; Johnson, 1984). Due to these rapid
sedimentation rates, reservoir sediments are considered
as being little aected by early diagenesis processes
(Callender, 2000) and provide preserved historical heavy
metal inputs.
Reservoir sediments are also a great concern since
they can turn from sink to source of heavy metals for
uvial systems by diusion at the watersediment
interface (e.g. Van Den Berg et al., 1999; Blanc et al.,
1999; Blasco et al., 2000), bioturbation (e.g. Forster
et al., 1999; Zoumis et al., 2001) or resuspension due to
dredging (e.g. Simpson et al., 2000; Van Den Berg et al.,
2000; Audry et al., 2004b) or ooding. Thus, it is
important to determine the intensity of pollution by
inventorying the heavy metal concentrations and their
spatial distribution in sediments.
The purpose of this study was to describe the
magnitude and historic variation of the deposition of
metallic pollutants in order to quantify the intensity of
the heavy metal pollution. This study also aims to
reconstruct the impact of anthropogenic activity on the
Lot River watershed by correlating the heavy metal
concentrationdepth proles (Cd, Zn, Cu and Pb) with
historical changes of industrial activity within the
watershed.

2. Material and methods

2.1. Study area and sampling sites
The Lot River watershed (11,840 km2), tributary of
the Gironde River, is known for its chronic Cd pollution
resulting from mining and smelting activities since the
late 19th century (e.g. Latouche, 1988; Grousset et al.,
1999; Blanc et al., 1999; Schafer et al., 2002; Audry et al.,
2004a). The main point source of Cd had been identied
in the upper part of the Lot River (Boutier, 1981;
Latouche, 1992), where a small tributary (Riou-Mort
River; Fig. 1) drains a waste area of a now-abandoned
zinc ore manufacturing facility in the Riou-Mort
watershed (155 km2). This factory had been active from
1842 to 1987. The end of the smelting activity followed
an accidental heavy metal pollution incident in 1986.
Zinc was produced from ZnCO3 until 1922 by thermic
reduction and then from ZnS by electrolysis. About
35 Mt (unpublished data) of smelting wastes still rich
in Zn and Cd (10,00020,000 mg kg
1 Zn and w200
400 mg kg
1 Cd) were produced and deposited on the
carboniferous and cristallophyllian (micaschists) bedrock. In 1971, rst remediation work was undertaken on
the nearest waste deposits from the factory, by collecting
and treating the draining water. In 1987, this remediation work was completed by conning part of the
waste deposits in storage basins with underlying and
overlying mud. Nevertheless, large parts of the waste
deposits are still exposed to rain water and thus leached
and eroded.
The sediments investigated were collected in three
articial lakes behind hydroelectric dams (reservoirs)
along the Lot River (Fig. 1). (1) Marcenac, located

Fig. 1. Map of the study area and the sampling sites shown as open stars. Also shown, is the smelting waste area where the heavy metal pollution
originates.

S. Audry et al. / Environmental Pollution 132 (2004) 413426

upstream of the conuence with the Riou-Mort River, in

order to obtain heavy metal background values for the
upper Lot River watershed. (2) Cajarc (0.3 km2;
established in 1950), the rst reservoir located 25 km
downstream of the conuence with the Riou-Mort
River. (3) Temple (3.8 km2; established in 1951),
upstream of the conuence of the Lot and Garonne
Rivers. This core is representative of sediment at the
outlet of the Lot River.

415

reference waters (SLRS-3, SLRS-4). Accuracy was

within 5% of the certied values and the analytical
error (relative standard deviation) generally better than
5% for concentrations 10 times higher than detection
limits.

3. Results
3.1. Description of the sediment cores

2.2. Sampling methods

In 2001, three continuous and w1.4-m-long sediment
cores were collected using a manual corer consisting of
a 10 ! 10 cm rectangular Plexiglas tube. The sediment
cores were sliced in thin horizontal sections with a plastic
cutter at 1 cm resolution from the surface to 25 cm for
Marcenac and Cajarc sites and at 2.5 cm resolution from
the surface to 10 cm for Temple site. The lower part of
the three cores were sampled with a 5 cm resolution.
Immediately after recovery, each sliced section was
centrifuged at 5000 rpm for 20 min in N2 ushed vials in
order to remove pore waters and then sealed in
sampling-bags under N2. For every level, an additional
sediment sub-sample was immediately sealed in a preweighed vial and frozen for porosity analyses. Back in
the laboratory, sediment samples were dried at 50  C to
constant weight and then powdered and homogenized
with an agate mortar.
2.3. Analytical methods
Grain size measurements were performed with a laser
granulometer (Malvern, Mastersizer type). Porosity was
determined by comparison of the weights of wet and
dried sediment. 137Cesium and 210Pb radionuclide
records were only measured on the sediment core
retrieved at the Cajarc site. These measurements were
done on 610 g dry samples with a low background
noise g-ray spectrometer (Intertechnique EGSP 220025), using a germanium detector. The precision of this
method is typically better than 10% (Jouanneau et al.,
1999).
Mineralization of sediment samples was undertaken
using clean techniques and labware following the
method previously described in detail in Schafer and
Blanc (2002) and Schafer et al. (2002). Each batch of
samples included method blanks and digestion of
certied international reference materials (e.g. CRM
320, PACS-1). Particulate heavy metal concentrations
were measured using ICP-MS (Elan 5000, Perkin
Elmer). Scandium concentrations were measured using
ICP-AES (Jobin Yvon 124) in the Centre de Geochimie de la Surface of Strasbourg (UMR CNRS
7517). The analytical methods employed were further
quality checked by analysis of certied international

The three cores consisted of dark, anoxic (dissolved

O2 saturation !0.5% at the watersediment interfaces;
data not shown), and unconsolidated sediment with
grain sizes ranging from ne sand to clay. The silty
fraction (!63 mm) represented w50% (Marcenac) and
w75% (Cajarc and Temple) of the whole sediment, with
the ne-silt fraction accounting for w25%. Very thin
layers of coarser sand (i.e. O250 mm) occurred in the
three sediment cores, probably representing markers of
ood events. Finally, little burrows were observed near
the watersediment interface. Radionuclide (137Cs and
210
Pb) activities were only measured in the Cajarc core.
Unfortunately, 210Pb measurements (data not shown)
could not be used to estimate the bioturbation eect on
the sediment as 210Pb activities in this core were likely to
be mainly controlled by grain size and by the
hydrological variability of the dierent sub-watersheds
of the Lot River (as discussed below). Nevertheless, the
impact of bioturbation eect on the sedimentary record
should be negligible because of the low density and size
of the burrows.
The results of porosity measurements are shown in
Fig. 2. For the Marcenac core, porosity averaged 80%
in the rst 35 cm, then dropped and stayed around 60%,
except at the depth interval centered around 80 cm that
showed porosity values near 80%. For the Cajarc core,
porosity sharply decreased from 89 to 62% within the
rst 15 cm and then remained around 70%. Finally, for
the Temple core, porosity decreased with increasing
depth from 81% at the watersediment interface to 40%
at 100 cm depth (Fig. 2).
3.2. Comparison of heavy metal concentrations with
natural background values and other river sites
Table 1 gives the maximum, minimum and mean
heavy metal concentrations of Cu, Zn, Cd and Pb in the
three reservoir sediments of the Lot River and in
sediments of other polluted sites. The Lot River
reference site (Marcenac) showed mean heavy metal
values clearly below those at the other sites listed in
Table 1. The mean values at Marcenac site were only
1.9, 2.6, 8.1 and 2.6 times higher for Cu, Zn, Cd and Pb,
respectively, than the mean UCC (upper continental
crust; Wedepohl, 1995; Table 1).

416

S. Audry et al. / Environmental Pollution 132 (2004) 413426

Cajarc
Porosity (%)

Marcenac
Porosity (%)
0

20 40 60 80 100

Temple
Porosity (%)

20 40 60 80 100

20 40 60 80 100

20

Depth (cm)

40
60
80
100
120
140
160
Fig. 2. Porosity (%) vs depth proles for sediment cores at Marcenac, Cajarc and Temple reservoir lake sites.

Compared to the reference site, sediments of Cajarc

site showed maximum Cd, Zn, Pb and Cu concentrations,
about 210, 50, 19 and 6 times higher, respectively. At
Temple site, the heavy metal concentrations were lower
compared to Cajarc site yet high with values 130, 35, 9 and
2 times higher compared to Marcenac site. These results
suggest severe pollution in the Lot River downstream
sediments (as discussed below). Indeed, heavy metal
concentrations in the Lot River downstream sediments
were highly elevated compared to average crustal concentrations (Table 1) with, for instance, concentrations
up to 2900-fold, 193-fold, 75-fold and 18-fold higher for
Cd, Zn, Pb and Cu respectively, at Cajarc site. Compared
to other sites aected by mine drainage listed in Table 1,
Zn and Cd maximum concentrations at Cajarc and
Temple sites were at least one order of magnitude
higher. Furthermore, Pb maximum concentration

(1280 mg kg
1) at Cajarc site was very close to that of

the Spanish polluted rivers. In contrast, the maximum
concentration of Cu (264 mg kg
1, Cajarc site) was about
10 times lower than that of sediments of Tinto and Odiel
Rivers but similar to that of the Guadiamar River
sediments (i.e. 265 mg kg
1; Table 1).
The lower heavy metal concentrations observed in the
Temple site sediments were very close to those measured
at the reference site. These minimum values occurred in
the bottom of the core (see Fig. 5) and can be interpreted
as sediment deposited before the beginning of the
industrial activity in the study area (i.e. 1842; Jozereau,
1997) or riverbank sediment. Thus, from these results,
natural background heavy metal values for the Lot
River watershed are proposed: w17 mg kg
1 (Cu),
w82 mg kg
1 (Zn), w0.33 mg kg
1 (Cd) and w28 mg
kg
1 (Pb).

Table 1
Maximum Cu, Zn, Cd and Pb concentrations in reservoir sediments from the Lot River, other selected sites and UCC (upper continental crust)
Location

Cajarc site, Lot River

Temple site, Lot River
Marcenac site, Lot River
UCC
Malter Reservoir, Germanya
Odiel River, Spaina
Tinto River, Spaina
Guadiamar River, Spaina
Patroon Reservoir, NY, USA
Lochnagar, Scotland
Lake Zurich, Switzerland
Meuse River, Netherlands
a

River draining mine sites.

Cu

Zn

Cd

Pb

Reference

Min. Max. Mean Min. Max.

Mean Min. Max. Mean Min. Max. Mean

40.6
16.7
17.9

54
180
98

8
20
50

4430
949
134
52
1362
954
1754
1421

220
912

264
78
42

240
2920
2650
265

25
78
105

97.7
30.7
26.9
14.3
196
1282
846
165

48.9
66

909
81.5
82.3

143
110
590

39
50
803

10,000
7230
206

1900
3150
6730
2970

180
675
1083

12.6
0.56
0.33

0.35
!1
1.8

0.3
0.2
8.2

294
180
1.4

91
68
23
9.5
25
1.9
19
11

125
20.4
0.81
0.10
37.5
8.5
6.2
3.9

2.1
9.8

111
31.3
27.9

46
370
51

100
10
121

1280
593
68

740
1140
1650
250
3600
360
150
232

523
105
43.6
17
465
649
870
127

94.4
135

This study
This study
This study
Wedepohl (1995)
Muller et al. (2000)
Morillo et al. (2002)
Galan et al. (2003)
Lacal et al. (2003)
Arnason and Fletcher (2003)
Yang et al. (2002)
Von Gunten et al. (1997)
Van Den Berg et al. (1999)

417

S. Audry et al. / Environmental Pollution 132 (2004) 413426

4. Discussion
4.1.

137

137

Cs dating and sedimentation rates

25

Cs (Bq kg-1)
50

75

100

20

2.8 cm a

40

1986

60

-1

2.4 cm a

1996

1986

-1

1976

80
1966
100

120

140

1963

estimated time scale (year)

Lead (t1/2 = 22.3 a) has been used for sediment

dating over a time scale of the last 100150 a. The
combination of 210Pb with 137Cs is often used to reduce
the uncertainty in dating (e.g. Benninger et al., 1997).
Typically, 210Pb dating of river sediment has shown
good results for rivers draining a small watershed (e.g.
Farmer et al., 1997; Eades et al., 1997). Measurements
of 210Pb activities in the Cajarc reservoir sediment (data
not shown) were performed in the bulk sediment and
appeared to be inuenced by grain size. This result is not
surprising since adsorption of 210Pb is known to be
greater for ne particles (e.g. Monna et al., 1996). In
addition, the 210Pb data were characterized by a great
variability with depth. It is likely that the variability of
the data could be related to the dierences in lithology,
hydrology and response-time between the sub-watersheds of the Lot River watershed upstream of Cajarc
(w8000 km2). These watersheds integrate dierent
lithologies with the respective 210Pb activities and
particle and water retention-times that may explain the
variability of the 210Pb data in the studied river
sediments. Consequently, it appeared that 210Pb activities could not provide reliable sediment dating.
137
Cesium (t1/2 = 30.2 a) found in marine and uvial
sediments is an anthropogenic radionuclide originating
from above-ground testing of nuclear explosives carried
out from 1945 to 1972, but 90% of the radionuclide
releases occurred between 1963 and 1964 (Ritchie and
McHenry, 1990; Callender and Robbins, 1993). 137Cesium has been used as an indicator for the amount of
global fallout deposited in soil (Beck and Krey, 1983)
and reservoir sediments (Krey et al., 1990). The use of
137
Cs as an indicator of sedimentation processes is
consistent as it binds almost irreversibly to clay and silt
particles (e.g. Francis and Brinkley, 1976; Poinssot et al.,
1999) and because of its long half-life (30.2 years).
Moreover, the Chernobyl nuclear accident of 1986 has
been recorded by the European sediments (Grousset
et al., 1999; Muller et al., 2000; Smith et al., 2000a).
Thus, 137Cs activity depth proles are often used for
sediment core dating (e.g. Grousset et al., 1999; Muller
et al., 2000; Jha et al., 2003; Abril, 2004).
The plot of 137Cs activity against depth for the Cajarc
sediment core is displayed in Fig. 3. Two peaks of 137Cs
activity are identied at 42.5 and 97.5 cm depth with
corresponding activities of 60 and 65 Bq kg
1. Uncertainties in the interpretation of 137Cs depth prole may
arise from possible post-depositional processes related
to 137Cs mobility in the sediment. Although 137Cs is
known to strongly bind mainly to the clay lattice, it can
also be sorbed exchangeably to the sediment (Comans,
1998; Smith et al., 2000b). It was further shown by

Depth (cm)

210

1956

1946

137

Fig. 3. Cs activity vs depth prole for the sediment core at Cajarc

site. Also shown, are the deduced mean sedimentation rates (see text
for details).

Smith and Comans (1996) that Cs may desorb from the

xed fraction with a half-time of approximately 10 a.
Typically, 137Cs depth proles aected by signicant Cs
post-depositional remobilisation show a tail of activity
penetrating several centimeters below the peak (e.g.
Meili and Worman, 1996). The 137Cs depth proles in
the sediments of Cajarc reservoir did not show obvious
tailing below both peaks, indicating that 137Cs remobilisation was negligible. This result is in agreement with
the conclusions of Smith et al. (2000b) stating that in
most of the lakes, the majority of 137Cs is strongly
sorbed to the sediment. The small shoulder in 137Cs
activities located at 3038 cm (Fig. 3) is attributed to
higher particle inputs, possibly related to ood events, as
suggested by the coarser median grain size (i.e. 44.7 mm)
of the plateau compared to the median grain size of the
shallower part of the core (34 mm). A similar plateau in
the 137Cs activity depth prole has already been
explained by high discharge events in Hudson River
sediments (Chillrud et al., 2003). Considering the overall
well-resolved 137Cs depth prole (i.e. without downward
tailing), migration of the major 137Cs horizons due to
post-depositional processes can be excluded.
The deepest 137Cs peak is assigned to the maximum
atmospheric global fallout that corresponded to 1963.
The shallowest 137Cs peak is attributed to the Chernobyl
accident of 1986. Mean sedimentation rates were
estimated using the 137Cs peaks of 1963 and 1986 by
dividing the sediment depth where the peak is located by
the proper time elapsed. Thus, mean sedimentation rates
of 2.4 cm a
1 were calculated between 1963 and 1986 and
2.8 cm a
1 between 1986 and 2001 (Fig. 3). An estimated

418

S. Audry et al. / Environmental Pollution 132 (2004) 413426

time scale was then deduced from the mean sedimentation rate values (Fig. 3). Weaker sediment compaction
(i.e. greater porosity, Fig. 2) of the rst 40 cm compared
to the rest of the core could explain the greater
sedimentation rate calculated for the 19862001 period.
The mean sedimentation rates calculated in this study are
of the same magnitude as those usually found for
reservoirs (e.g. 2.9 cm a
1, Muller et al., 2000; 4 cm a
1,
Arnason and Fletcher, 2003). They are much higher than
those found for natural lakes (i.e. !1 cm a
1, Hakanson
and Jansson, 1983; !0.3 cm a
1, Johnson, 1984; Von
Gunten et al., 1997). Thus, these high sedimentation
rates associated with sediment composed mainly of siltyclay particles are well-suited environmental conditions
for sedimentary recording of the heavy metal pollution.
4.2. Heavy metal concentration proles
4.2.1. Total heavy metal concentrations
In this study, V proles are presented as representative of the behavior of heavy metals (Cr, Co, Ni, Th, U;
data not shown) that were not proved to be enriched in
the studied sediments. The reference core at Marcenac
site showed very similar Cu, Zn, Cd, Pb and V proles
(Fig. 4a). The rst 20 cm showed variable heavy metal
concentrations, with variability exceeding 50%. Below
20 cm depth, the Cu, Zn, Cd, Pb and V proles were
more or less constant, centered on concentrations of 20,
100, 0.5, 30 and 50 mg kg
1, respectively.
At Cajarc site, the rst 10 cm of the sediment core
showed nearly constant Cu, Zn, Cd, Pb and V concentrations of 4149, w1000, w17, w120 and 71
79 mg kg
1, respectively (Fig. 4b). Below 10 cm depth,
Cu, Zn, Cd and Pb, showed similar proles characterized
by an increase in the average concentrations reaching
values 2.5-fold (Cu), 5-fold (Pb), 7-fold (Zn) and 9-fold
(Cd) higher than those at the watersediment interface.
Within the 42.545 cm depth interval, Cu, Zn, Cd and Pb
concentrations increased drastically reaching 153, 10,000,
300 and 930 mg kg
1, respectively. This depth interval
also showed the 137Cs activity peak (see above). Thus,
these heavy metal concentration peaks are interpreted
resulting from the accidental heavy metal pollution that
occurred in the Lot River watershed in 1986 (see above).
Below 45 cm depth, the heavy metal proles were highly
variable and showed high concentrations compared to
the surface sediment (Fig. 4b). In contrast to the other
metals, V proles did not show important peaks either at
42.545 cm depth or in the lower part of the core. Indeed,
variability was generally low.
At Temple site, Cu, Zn, Cd and Pb proles showed
concentrations of 35, 620, 11 and 80 mg kg
1 at the
watersediment interface, respectively (Fig. 4c). Below,
these proles were characterized by two concentration
peaks roughly of the same amplitude at 35 and 50 cm
depth, corresponding to concentrations 1.2-fold (Cu),

4.8-fold (Zn), 6.4-fold (Cd) and 2.2-fold (Pb) higher

than the values at the watersediment interface. Lead
prole showed an additional concentration peak at
68.5 cm depth of about the same amplitude as the two
above ones. Downward, heavy metal concentrations
declined rapidly. Consequently, at the bottom of the
core, heavy metal proles showed concentrations (16.7,
81.5, 0.56 and 31.2 mg kg
1 for Cu, Zn, Cd and Pb,
respectively) much lower than those at the water
sediment interface (Fig. 4c). Additionally, these bottom-sediment heavy metal concentrations were similar
to the minimum concentrations characterizing the sediments of the Marcenac reservoir (i.e. 17.9, 82.3, 0.33 and
27.9 mg kg
1, respectively; Table 1). Finally, V concentrations showed a very dierent prole compared to the
other heavy metal proles with decreasing values from
the interface (95.4 mg kg
1) to 38.5 cm depth
(44 mg kg
1). Below this depth, V concentrations were
relatively constant, i.e. 58.3 G 1.5 mg kg
1.
As 137Cs activity was not measured at Temple site,
estimates of the timing of heavy metal deposition and
average sedimentation rates were done by comparing the
heavy metal concentrationdepth proles at Cajarc and
Temple sites. Although one could expect heavy metal
records at Cajarc and Temple sites to be roughly identical, the concentrationdepth proles at both sites were
very dierent. The bottom heavy metal concentrations
at the Temple core were too low to be coherent with the
historic Zn production in the Riou-Mort watershed.
Thus, the lower part of the core can be interpreted as old
riverbed sediment or as former soil (before dam
construction and establishment of the smelting factory).
This interpretation of the data is consistent with the fact
that the low heavy metal concentrations in the bottomsediment of the Temple core were very close to those in
the Marcenac core (see above). Dating the bottom of the
Temple site core at 1842 and the top of the core at 2001
(coring date) leads to a mean sedimentation rate of
w0.6 cm a
1. This value is much lower than that
calculated at Cajarc site (i.e. 2.42.8 cm a
1; Fig. 3)
and thus is considered to be greatly underestimated.
Moreover, the heavy metal peak observed at 33.5 cm
depth at Temple site can be reasonably attributed to the
accidental pollution of 1986. On this basis, we estimate
an average sedimentation rate of w2.3 cm a
1 for the
period between 1986 and 2001, that is close to that
calculated at Cajarc site. Thus, we propose that the
sedimentary record at Temple site was probably incomplete, showing a sedimentary hiatus before 1986
possibly due to dam ushing that could have been
responsible for high sediment remobilisation.
4.2.2. Normalization of the heavy metal
proles to scandium
In order to dierentiate between lithogenic and
anthropogenic origins of heavy metal in the Lot River

419

S. Audry et al. / Environmental Pollution 132 (2004) 413426

(a)

Cu (m g kg-1)
0

25

Cd (m g kg-1)

Zn (m g kg-1)
50

125

250

Pb (m g kg-1)
2

25

50

V (m g kg-1)
75

40

80

120

20

De pth ( cm )

40
60
80
100
120
140

(b)

Cu (m g kg-1)
0

120

Zn (m g kg-1)
240 0

6 000

Pb (m g kg-1)

Cd (m g kg-1)

12 000 0

200

400

650

V (m g kg-1)

1 300 0

50

100 150

20

De pth ( cm )

40

1986

60
80
1963

100
120
140

(c)

Cu (m g kg-1)
0

25

50

75

Zn (m g kg-1)

Cd (m g kg-1)

2 000

50

4 000 0

Pb (m g kg-1)
100

150

V (m g kg-1)
300

50

100 150

De pth ( cm )

20
40
60
80
100

Fig. 4. Depthconcentration proles for Cu, Zn, Cd, Pb and V for the Marcenac (a), Cajarc (b) and Temple (c) cores. Dashed horizontal lines
correspond to the date markers of 1963 and 1986 from 137Cs data.

reservoir sediments, the heavy metal proles were

normalized to Sc. Scandium is a conservative (very
insoluble) element hosted in a variety of ferromagnesian
minerals (Frondl, 1970). The reason for choosing Sc
instead of other conservative elements like Ti or Zr is its
wide distribution in crustal rocks, unlike Ti and Zr
which are mainly concentrated in ilmenite and zircon,
respectively (Steinmann and Shotyk, 1997).

The results of the normalization of heavy metal

concentrations to Sc are shown in Fig. 5. At Marcenac
site, heavy metals did not show any signicant
enrichment along the whole core with at Sc-normalized
proles. At Cajarc and Temple sites, Sc-normalized
proles of Cu, Zn, Cd and Pb were very similar to the
non-normalized proles, i.e. showing similar heavy
metal variations. In contrast, V and other heavy metals

420

S. Audry et al. / Environmental Pollution 132 (2004) 413426

(a)
0

Cu/Sc
2
4

Zn/Sc
10

Cu/Sc
10

20

Zn/Sc
1 000

20

Cd/Sc
0.1

0.2

Cd/Sc
10 20

30

10

V/Sc
5

10

Pb/Sc
50 100 150

V/Sc
10

20

Pb/Sc
5

20

Depth (cm)

40
60
80
100
120
140

(b)
0
20

Depth (cm)

40

1986

60
80
1963

100
120
140

(c)
0

Cu/Sc
2
4

Zn/Sc
200

400

Cd/Sc
5

10

Pb/Sc
15

30

V/Sc
5

10

Depth (cm)

20
40
60
80
100

Fig. 5. Scandium-normalized proles of Cu, Zn, Cd, Pb and V concentrations for the Marcenac (a), Cajarc (b) and Temple (c) cores.

(Cr, Co, Ni, Th, U, data not shown) showed at proles

for these two sites. These results indicate clearly the
great dominance of anthropogenic over lithogenic
sources for Cu, Zn, Cd and Pb inputs aecting the Lot
River sediments downstream of the conuence between
the Lot and the Riou-Mort Rivers. On the other side,
inputs of heavy metals such as V, Cr, Co, Ni, Th and U
did not seem to be dominated by anthropogenic/point
sources and thus originated mainly from lithogenic/
diuse sources.

4.3. Intensity of the heavy metal pollution

In order to evaluate the extent of the historical heavy
metal pollution aecting the Lot River downstream sediments, an enrichment factor (EF) is calculated as follows:
EF Me=Scsed =Me=ScMarcenac

where (Me/Sc)sed is the concentration ratio of a given

metal to Sc in the Lot River sediment and (Me/

421

S. Audry et al. / Environmental Pollution 132 (2004) 413426

Sc)Marcenac is the average ratio of the same metal to Sc in

the reference (non-polluted) site. Quite often, it is
dicult to determine the typical background Me/Sc
ratio for a given watershed due to the complex lithology
and weathering features within the drainage areas
(Zhang, 1995). Consequently, it is essential to choose
appropriate samples with Me/Sc ratios representative of
the non-polluted part of the watershed. Thus, as
expressed by Eq. (1), we consider the average Me/Sc
ratio in sediment from Marcenac site as a reference for
the evaluation of EF related to anthropogenic point
source in the Riou-Mort watershed. Depth proles of
EF at Cajarc site for Cu, Zn, Cd and Pb generally
showed values above 1 (Fig. 6a), conrming the
pollution of the Lot River downstream sediment.
Cadmium showed the heaviest pollution with EF up to
300 (42.5 cm depth) that were always much higher than
those of Cu, Zn and Pb. The EF values along the whole
Cajarc core indicate that the accidental pollution of
1986 (42.5 cm depth; see above) was the major pollution
event recorded in the reservoir sediments of the Lot
River since the 1950s. Nevertheless, other very intense
Cd enrichments characterized the downward part of the
core, particularly from the core bottom to 82 cm depth
(i.e. 1969) indicating very high pollution, probably due
to intensication after World War II. The EF values in
reservoir sediments from Temple core were lower than

(a)
0

50

Enrichment Factor
100
150
200

Depth (cm)

Igeo log2

250

where [Me]sed is the concentration of a given heavy

metal and [Me]Marcenac is the average concentration of
the same metal in the sediments of the reference site. The
Igeo is associated with a qualitative scale of pollution
intensity (Table 2). Following this classication, the
sediments at Cajarc site, with Igeo values above 5 from
the bottom of the core to 17 cm depth (i.e. 19501997),
are very strongly polluted by Cd. Accordingly, the
sediments are strongly polluted by Zn (3 ! Igeo O 4),
moderately to strongly polluted by Pb (Igeo O 2) and
moderately polluted by Cu (Fig. 6c). Since 1997,
despite the remediation eorts undertaken in the

Igeo

300

10

0
20

40

1986

60

1986

40
60

80

80
1963

100

120

140

140

(b)
50

100

150

200

Igeo

(d)

Enrichment Factor
0

1963

100

120

Mesed
1:5MeMarcenac

(c)

20

Depth (cm)

those of Cajarc sediments with maximum values up to

98, 25 and 5.9 for Cd, Zn and Pb, respectively. The lesser
extent of heavy metal pollution at Temple site compared
to Cajarc site is attributed to the downstream dilution of
the heavy metal concentration by the water and
suspended material from the Cele River, a tributary of
the Lot River (Fig. 1) and direct erosion inputs.
A second criterion to evaluate the intensity of the
historical heavy metal pollution in the Lot River
reservoir sediments is the geoaccumulation index (Igeo)
introduced by Muller (1979):

250

-2

300

10

20

20

40

40

60

60

80

80
100

100

Cu

Pb

Cd

Zn

Fig. 6. Enrichment factor (EF) and geoaccumulation index (Igeo) for Cu, Zn, Cd and Pb in the sediment cores of Cajarc (a,c) and Temple (b,d) sites.

422

S. Audry et al. / Environmental Pollution 132 (2004) 413426

Table 2
Geoaccumulation index (Igeo) in relation to pollution intensity after
Muller (1979)
Igeo

Pollution intensity

O5
45
34
23
12
01
!0

very strongly polluted

strongly to very strongly polluted
strongly polluted
moderately to strongly polluted
moderately polluted
unpolluted to moderately polluted
unpolluted

Riou-Mort watershed, the levels of pollution by Cd and

Zn are still high with Igeo values around 4 and greater
than 2, respectively. Downstream, at Temple site, heavy
metal pollution is less intense, only Cd and Zn showing
maximum Igeo values above 4.
Thus, using the Igeo classication, the intensity of Lot
River sediment pollution by heavy metal follows the
relative order: Cd Z Zn O Pb O Cu. Furthermore, Lot
River sediments contain Cd, Zn, Pb and Cu concentration levels that could be severely detrimental to aquatic
biota (e.g. NYSDEC, 1994). Indeed, sediment resuspension during oods or due to anthropogenic activities (i.e.
dredging) will remobilise reduced particulate heavy
metals that may be oxidized and, thus, become
potentially available for biota (e.g., Audry et al., 2004b).
4.4. Estimate of the heavy metal inventory
within the Lot River sediments
Despite the variability of Zn and Cd concentrations
in the sediment cores (see Fig. 4), the Zn/Cd ratio was
nearly constant within each core but clearly dierent
between the three cores (Fig. 7). The reference site
sediments (Marcenac) were characterized by high Zn/Cd
ratios (i.e. 169 G 25). In contrast, suspended particulate
matter (SPM) in the Riou-Mort River (data from Audry
et al., 2004b) were characterized by very low Zn/Cd
ratios (i.e. 45 G 7; Fig. 7), conrming high Cd enrichment of the material transported by the Riou-Mort
River to the Lot River. The mean Zn/Cd ratio of 47 G 6
(Fig. 7a) in the Cajarc sediments was very close to that
of the Riou-Mort SPM. The latter result clearly
indicates the great inuence of the Riou-Mort River
on the Lot River heavy metal budgets although the
Riou-Mort River only contributes to less than 1% of the
Lot River discharge (Blanc et al., 1999). Nevertheless,
since 1997 (i.e. the upper 12 cm of the core), the Cajarc
site sediments are characterized by slightly higher Zn/Cd
values (i.e. 63 G 12; Fig. 7a), reecting decreased Cd
inputs derived from the anthropogenic source related to
the remediation eorts undertaken in the Riou-Mort
watershed (see above). These results are in agreement
with the decrease in the total Cd uxes since 1996 in the
overall Lot River watershed (Audry et al., 2004a).

Fig. 7. Zn/Cd ratio (mean G standard deviation) vs Cd concentrations

in the sediments of Marcenac, Cajarc (a), Temple (b) sites and in the
Riou-Mort River particulate suspended matter (SPM; *data from
Audry et al., 2004b).

At Temple site, upper sediments showed Zn/Cd ratios

(i.e. 50 G 7; Fig. 7b) close to those at the Cajarc site,
whereas the last 15 cm of bottom-sediment were
characterized by lower Zn/Cd ratio values (i.e.
115 G 24). This result conrms a deposition time for
these sediments previous to the beginning of the
industrial activity responsible for the heavy metal
pollution (see above).
Using Zn/Cd, Pb/Cd and Cu/Cd ratios in the Cajarc
and Temple sediments (Table 3) and the estimate of the
Cd inventory in the Lot River sediments (i.e. 207 G 42 t;
Lapaquellerie et al., 1995), we estimated the inventory of
Zn, Pb and Cu in the Lot River sediments (Table 4). The
average Zn/Cd and Pb/Cd ratios of the sediments of
Cajarc and Temple sites were nearly similar, whereas the
mean Cu/Cd ratio was higher at the downstream site (1.1
at Cajarc and 1.8 at Temple; Table 3). This increase can
be attributed to additional Cu sources in the downstream valley of the Lot River, e.g. erosion of vineyards
treated with CuSO4 and/or Cu ore deposits (BRGM,
1978; Schafer and Blanc, 2002). The estimated heavy
metal inventory suggests that about 10,000, 1000 and
300 t of Zn, Pb and Cu, respectively, are stocked in the
Lot River sediments. Moreover, the results showed that
w30% of the Cd, Zn, Cu and Pb inventory are trapped
in the downstream sediments of the Temple reservoir,
near the conuence between the Lot and the Garonne
Rivers. The latter result is an important issue since the

423

S. Audry et al. / Environmental Pollution 132 (2004) 413426

Table 3
Mean values of the Zn/Cd, Pb/Cd and Cu/Cd ratios in the sediments
of the Cajarc and Temple reservoir lakes and average for the Lot River

Cajarc
Temple
Lot River
a

Area
(km2)

Sediment
volumea
(.106m3)

Zn/Cd

0.35
3.26
w26

0.12
3.92
9.47

47 G 6
50 G 7
48.5 G 6.5

Pb/Cd

Cu/Cd

Table 4
Inventory of Cd, Zn, Pb and Cu in the sediments of Cajarc and Temple
reservoir lakes and in the Lot River sediments downstream of the
conuence of the Riou-Mort and Lot Rivers
Heavy metal inventory (t)

4.8 G 2.1
4.9 G 1.2
4.85 G 1.6

1.1 G 0.8
1.8 G 1.2
1.45 G 1.0

Data from Lapaquellerie et al. (1995).

dam gates are episodically opened during ood events

and thus might release heavy metal-polluted sediments
into the Garonne River.
4.5. Relation between the heavy metal proles
and historical changes in the smelting activity
in the Riou-Mort River watershed
On the basis of the 137Cs dating and the heavy metal
concentrationdepth proles of the sediments of Cajarc
reservoir, we propose to connect the sedimentary record

Cajarc
Temple
Lot River
a

Cda

Zn

Pb

Cu

23.1
55
207 G 42

1090 G 138
2750 G 385
10,040 G 1448

111 G 45
264 G 116
1000 G 341

25.4 G 18.5
99 G 49
300 G 207

Data from Lapaquellerie et al. (1995).

of the heavy metal pollution with known historical

evolution of production and waste-treatment proceedings in the industrial area of the Riou-Mort watershed.
For this, we chose the Sc-normalized sedimentary Cd
and Cu records (Fig. 8).
From the bottom of the core to 65 cm depth
(w1977), heavy metal levels were high and very variable
indicating about three decades of high heavy metal
inputs from the Riou-Mort watershed to the Lot River.
From 1977 to 1985 (6545 cm depth), the sedimentary
record showed an increase of Cd levels indicating clearly

Fig. 8. Relation between heavy metal sedimentary record (Cd/Sc and Cu/Sc) in the reservoir lake sediments of Cajarc and the historical evolution of
the treatment proceedings of the smelting activity-related wastes. *Data from Jozereau (1997).

424

S. Audry et al. / Environmental Pollution 132 (2004) 413426

increased heavy metal inputs. This can be reasonably

attributed to the broadening of the smelting waste area
in the Riou-Mort watershed due to maximal smelting
activity (i.e. Zn production = 100,000 t a
1; Jozereau,
1997) during this period of time. Nevertheless, Cu levels
clearly showed a decrease simultaneously with the Cd
increase. The antagonistic evolution of both metal levels
can be explained by increased metal recoveries due to
modern production techniques (i.e. electrolysis; Fig. 8)
more ecient for Cu than for Cd and Zn.
The accidental pollution of 1986 (42.5 cm depth) was
well recorded in the sediments that showed maximum
heavy metal concentrations at the respective depth
range. The steep decrease of heavy metal concentrations
in the upper part of the core is attributed to the
remediation eorts (i.e. draining water treatment and
conning smelting waste) undertaken since 1987. Nevertheless, the more or less constant heavy metal levels
since 1997 (upper 15 cm of the core) are still much
higher than those at the upstream reference site (i.e. 13fold, 6-fold, 2-fold and 1.2-fold higher for Cd, Zn, Pb
and Cu, respectively; see Fig. 4). This indicates clearly
that despite the remediation eorts undertaken since the
accidental pollution of 1986, the Lot River watershed is
still impacted by anthropogenic heavy metal inputs from
the Riou-Mort watershed. This is attributed to continued weathering of the smelting-related waste area.
Consequently, the Lot River is still subjected to chronic
heavy metal pollution.
5. Conclusions
The present study of trace metal abundances and
distributions in 137Cs dated reservoir sediment cores
from the Lot River conrms the historic (O50 years)
polymetallic pollution due to former industrial activities
in the Riou-Mort watershed:
1. The Lot River reservoir sediments showed wellpreserved 137Cs peaks corresponding to maximum
atmospheric 137Cs inputs in 1963 (nuclear weapon
tests) and 1986 (Chernobyl accident).
2. High mean sedimentation rates of 2.42.8 cm a
1
were calculated for the sediments of the reservoir
located downstream of the anthropogenic heavy
metal inputs. These sediments clearly recorded the
temporal evolution of the heavy metal pollution
(Cu, Zn, Cd and Pb) within the Lot River
watershed.
3. The major polluting event that has impacted the
Lot River sediment during the recorded period
coincided with the Chernobyl accident.
4. Based on comparison between heavy metal concentrations in sediments upstream of the anthropogenic inputs and bottom-sediments of the furthest
downstream core (interpreted as old soil or

riverbed), concentrations of w17, w82, w0.33

and w28 mg kg
1 for Cu, Zn, Cd and Pb, respectively, are proposed as natural background values
for the Lot uvial system.
5. The Lot River sediments still contain important
heavy metal inventory: w10,000, w1000 and 300 t
for Zn, Pb and Cu, respectively.
6. The intensity of the heavy metal pollution was
evaluated by calculating enrichment factors that
showed maximum values of 300 and 65 for Cd and
Zn, respectively. Furthermore, according to the
geoaccumulation index (Igeo) classication, the Lot
River sediments must be considered as severely
polluted in Cd and Zn. The relative importance of
the pollution is as follows: Cd Z Zn O Pb O Cu.
Moreover, despite remediation eorts undertaken
at the former smelting site, the Lot River is
still severely (Igeo w 4) and moderately to
severely (Igeo O 2) impacted by Cd and Zn inputs,
respectively.

Acknowledgements
This study is a scientic contribution of the GIS
ECOBAG (Groupement dInteret Scientique: Ecologie du Bassin Adour-Garonne). The assistance of the
following colleagues covering various aspects of the eld
and laboratory work is gratefully acknowledged: Y.
Lapaquellerie, N. Maillet, G. Chaillou, P. Anschutz, G.
Lavaux from Bordeaux and J. Samuel and R. Rouault
from Strasbourg. The authors greatly appreciate the
thoughtful comments of Dr. J.M. Pates and an
anonymous reviewer during the review process. This
work was supported by grants from lAgence de lEau
Adour-Garonne, le Ministe`re de la Recherche and le
FEDER-Region Aquitaine.

References
Abril, J.M., 2004. Constraints on the use of 137Cs as a time-marker to
support CRS and SIT chronologies. Environ. Pollut. 129, 3137.
Arnason, J.G., Fletcher, B.A., 2003. A 40C year record of Cd, Hg, Pb,
and U deposition in sediments of Patroon Reservoir, Albany
County, NY, USA. Environ. Pollut. 123, 383391.
Audry, S., Blanc, G., Schafer, J., 2004a. Cadmium transport in the
LotGaronne River system (France)dtemporal variability and
a model for ux estimation. Sci. Total Environ. 319, 197213.
Audry, S., Schafer, J., Blanc, G., Bossy, C., Lavaux, G., 2004b.
Anthropogenic components of heavy metal (Cd, Zn, Cu, Pb)
budgets in the LotGaronne uvial system (France). Appl.
Geochem. 19, 269286.
Beck, H.L., Krey, P.W., 1983. Radiation exposure in Utah from
Nevada nuclear tests. Science 220, 1824.
Benninger, L.K., Suayah, I.B., Stanley, D.J., 1997. Manzala lagoon,
Nile delta, Egypt: modern sediment accumulation based on
radioactive tracers. Environ. Geol. 34, 183193.

S. Audry et al. / Environmental Pollution 132 (2004) 413426

Blanc, G., Lapaquellerie, Y., Maillet, N., Anschutz, P., 1999. A
cadmium budget for the LotGaronne uvial system (France).
Hydrobiologia 410, 331341.
Blasco, J., Saenz, V., Gomez-Parra, A., 2000. Heavy metal uxes at the
sedimentwater interface of three coastal ecosystems from southwest of the Iberian Peninsula. Sci. Total Environ. 247, 189199.
Boutier, B., 1981. Synthe`se des resultats de la surveillance des
micropolluants dans la matie`re vivante. Ministe`re de lEnvironnement. Bulletin du Reseau National dObservation 17, 115174.
BRGMdBureau de Recherches Geologiques et Minie`res, 1978. Cartes
des g tes mineraux de la France 1/500 000, feuille Lyon. Service
Geologique National.
Callender, E., Robbins, J.A., 1993. Transport and accumulation of
radionuclides and stable elements in a Missouri River reservoir.
Water Resour. Res. 29, 17871804.
Callender, E., 2000. Geochemical eects of rapid sedimentation in
aquatic systems: minimal diagenesis and the preservation of
historical metal signatures. J. Paleolimnol. 23, 243260.
Chillrud, S.N., Hemming, S., Shuster, E.L., Simpson, H.J., Bopp,
R.F., Ross, J.M., Pederson, D.C., Chaky, D.A., Tolley, L.-R.,
Estabrooks, F., 2003. Stable lead isotopes, contaminant metals and
radionuclides in upper Hudson River sediment cores: implications
for improved time stratigraphy and transport processes. Chem.
Geol. 199, 5370.
Comans, R.N.J., 1998. Kinetics and reversibility of radiocaesium
sorption on illite and sediments containing illite. In: Sparks, D.L.,
Grundl, T. (Eds.), Kinetics and Mechanisms of Reactions at the
Mineral/Water Interface. ACS Symp. Ser. 715, 179201.
Eades, L.J., Farmer, J.G., MacKenzie, A.B., Kirika, A., Bailey-Watts,
A.E., 1997. High-resolution prole of radiocaesium deposition in
Loch Lomond sediments. J. Environ. Radioact. 39, 107115.
Farmer, J.G., MacKenzie, A.B., Eades, L.J., Kirika, A., Bailey-Watts,
1997. Inuences on the extent and record of heavy metal pollution
in sediment cores from Loch Tay in a mineralised area of Scotland.
J. Geochem. Explor. 58, 195202.
Francis, C.W., Brinkley, F.S., 1976. Preferential adsorption of 137Cs to
micaceous minerals in contaminated freshwater sediment. Nature
260, 511513.
Forster, S., Glud, R.N., Gundersen, J.K., Huettel, M., 1999. In situ
study of bromide tracer and oxygen ux in coastal sediments.
Estuar. Coast. Shelf Sci. 49, 813827.
Frondl, C., 1970. Scandium. In: Wedepohl, K.H. (Ed.), Handbook of
Geochemistry. Springer-Verlag, Berlin.
Galan, E., Gomez-Ariza, J.L., Gonzalez, I., Fernandez-Caliani, J.C.,
Morales, E., Giraldez, I., 2003. Heavy metal partitioning in river
sediments severely polluted by acid mine drainage in the Iberian
Pyrite Belt. Appl. Geochem. 18, 409421.
Grousset, F., Jouanneau, J.M., Castaing, P., Lavaux, G., Latouche,
C., 1999. A 70 year record of contamination from industrial
activity along the Garonne River and its tributaries (SW France).
Estuar. Coast. Shelf Sci. 48, 401414.
Hakanson, L., Jansson, M., 1983. Principles of Lakes Sedimentology.
Springer-Verlag, Berlin, Germany, 317 pp.
Hochella Jr., M.F., Moore, J.N., Golla, U., Putnis, A., 1999. A TEM
study of samples from acid mine drainage systems: Metal-mineral
association with implications for transport. Geochim. Cosmochim.
Acta 19-20, 33953406.
Jha, S.K., Chavan, S.B., Pandit, G.G., Sadasivan, S., 2003. Geochronology of Pb and Hg pollution in a coastal marine
environment using global fallout. J. Environ. Radioact. 69, 145
157.
Johnson, T.C., 1984. Sedimentation in large lakes. Annu. Rev. Earth
Planet. Sci. 12, 179204.
Jouanneau, J.M., Castaing, P., Grousset, F., Buat-Menard, P.,
Pedemay, Ph., 1999. Recording and chronology of a cadmium
contamination by 137Cs in the Gironde estuary (SW France).

425

Comptes Rendus de Lacademie des Sciences, Paris, Earth Planet.

Sci. 329, 265270.
Jozereau, C., 1997. La production de zinc dans la vallee du Lot depuis
le XIXe`me sie`cle. Consequences sur lenvironnement. Technical
report. Master Diploma of University of Bordeaux I, 21 pp.
Kahkonen, M.A., Suominen, K.P., Manninen, P.K., SalkinojaSalonen, M.S., 1998. 100 years of sediment accumulation history
of organic halogens and heavy metals in recipient and nonrecipient
lakes of pulping industry in Finland. Environ. Sci. Technol. 32,
17411746.
Krey, P.W., Heit, M., Miller, K.M., 1990. Radioactive fallout
reconstruction from contemporary measurements of reservoir
sediments. Health Phys. 59, 541554.
Lacal, J., Pilar da Silva, M., Garc a, R., Sevilla, T.M., Procopio, J.R.,
Hernadez, L., 2003. Study of fractionation and potential mobility
of metal in sludge from pyrite mining and aected river sediments:
changes in mobility over time and use of articial ageing as
a tool in environmental impact assessment. Environ. Pollut. 124,
291305.
Lapaquellerie, Y., Jouanneau, J.M., Maillet, N., Latouche, C., 1995.
Cadmium pollution in the sediments of the Lot River (France).
Estimate of the mass of cadmium. Environ. Technol. 16, 1145
1154.
Latouche, C., 1988. Cadmium pollution in the Gironde estuary.
Bulletin de linstitut Geologique du Bassin dAquitaine, Bordeaux
44, 1521.
Latouche, C., 1992. La pollution par le cadmium des hu tres sauvages
de lEstuaire de la Gironde. Origine. Mecanismes responsables de
la xation du cadmium. Ichtyophysiol. Acta 15, 139152.
Monna, F., Lancelot, J., Bernat, M., Mercadier, H., 1996. Sedimentation rate in the Thau basin (France) according to geochronological, geochemical and stratigraphical data. Oceanol. Acta 20,
627638.
Meili, M., Worman, A., 1996. Desorption and diusion of episodic
pollutants in sediments: a 3-phase model applied to Chernobyl
137
Cs. Appl. Geochem. 11, 311316.
Morillo, J., Usero, J., Gracia, I., 2002. Partitioning of metals in
sediments from the Odiel River (Spain). Environ. Int. 28, 263271.
Muller, G., 1979. Schwermetalle in den Sedimenten des RheinsVeranderungen seit. Umschav 79, 133149.
Muller, J., Ruppert, H., Muramatsu, Y., Schneider, J., 2000. Reservoir
sedimentsda witness of mining and industrial development
(Malter Reservoir, eastern Erzgebirge, Germany). Environ. Geol.
39, 13411351.
New York State Department of Environmental Conservation (NYSDEC), 1994. Technical guidance for screening contaminated
sediments: Albany, NY, Division of Fish and Wildlife, Divison
of Marine Resources, 36 pp.
Poinssot, C., Baeyens, B., Bradbury, M., 1999. Experimental and
modelling studies on caesium sorption on illite. Geochim.
Cosmochim. Acta 63, 32173227.
Ritchie, J.C., McHenry, J.R., 1990. Application of radioactive fallout
137
cesium for measuring soil erosion and sediment accumulation
rates and patterns: a review. J. Environ. Qual. 19, 215233.
Salomons, W., 1995. Environmental impact of metals derived from
mining activities: Processes, predictions, prevention. J. Geochem.
Explor. 52, 523.
Schafer, J., Blanc, G., 2002. Relationship between ore deposits in river
catchments and geochemistry of suspended particulate matter from
six rivers in south-west France. Sci. Total Environ. 298, 103118.
Schafer, J., Blanc, G., Lapaquellerie, Y., Maillet, N., Maneux, E.,
Etcheber, E., 2002. Ten-year-observation of the Gironde tributary
uvial system: uxes of suspended matter, particulate organic
carbon and cadmium. Mar. Chem. 79, 229242.
Simpson, S.L., Apte, S.C., Batley, G.E., 2000. Eect of short-term
resuspension events on the oxidation of cadmium, lead, and zinc

426

S. Audry et al. / Environmental Pollution 132 (2004) 413426

sulde phases in anoxic estuarine sediments. Environ. Sci. Technol.

34, 45334537.
Smith, J.T., Comans, R.N.J., Beresford, N.A., Wright, S.M., Howard,
B.J., Camplins, W.C., 2000a. Chernobyls legacy in food and water.
Nature 405, 141.
Smith, J.T., Comans, R.N.J., 1996. Modelling the diusive transport
and remobilisation of 137Cs in sediments: the eects of sorption
kinetics and reversibility. Geochim. Cosmochim. Acta 60, 995
1004.
Smith, J.T., Comans, R.N.J., Ireland, D.G., Nolan, L., Hilton, J.,
2000b. Experimental and in situ study of radiocaesium transfer
across the sedimentwater interface and mobility in lake sediments.
Appl. Geochem. 15, 833848.
Steinmann, P., Shotyk, W., 1997. Geochemistry, mineralogy, and
geochemical mass balance on major elements in two peat bog
proles (Jura Mountains, Switzerland). Chem. Geol. 138, 2553.
Van Den Berg, G.A., Loch, J.P.G., Van Der Heijdt, L.M., Zwolsman,
J.J.G., 1999. Mobilisation of heavy metals in contaminated

sediments in the river Meuse, The Netherlands. Water Air Soil

Pollut. 166, 567586.
Van Den Berg, G.A., Meijers, G.G.A., Van Der Heijdt, L.M.,
Zwolsman, J.J.G., 2000. Dredging-related mobilisation of trace
metals: a case study in the Netherlands. Water Res. 35, 19791986.
Von Gunten, H.R., Sturm, M., Moser, R.N., 1997. 200-year record of
metals in lake sediments and natural background concentrations.
Environ. Sci. Technol. 31, 21932197.
Wedepohl, K.H., 1995. The composition of the continental crust.
Geochim. Cosmochim. Acta 59, 12171232.
Yang, H., Rose, N.L., Battarbee, R.W., 2002. Distribution of some
trace metals in Lochnagar, a Scottish mountain lake ecosystem and
its catchment. Sci. Total Environ. 285, 197208.
Zhang, J., 1995. Geochemistry of trace metals from Chinese River/
Estuary systems: an overview. Estuar. Coast. Shelf Sci. 41, 631658.
Zoumis, T., Schmidt, A., Grigorova, L., Calmano, W., 2001.
Contaminants in sediments: remobilisation and demobilisation.
Sci. Total Environ. 266, 195202.