STUDY OF PARTICULATE MATTER AND HEAVY-METAL FLUXES IN
THE VENICE CANAL NETWORK BY LARGE VOLUME FILTRATION AND SEDIMENT TRAPS
Luca
Zaggia, Roberto Zonta, and Flaviano Collavini (National Research Council -
I.S.D.G.M., S. Polo 1364, Venezia - 30125, Italy. e-mail:zaggia@isdgm.ve.cnr.it);
Cinzia Bettiol (Department of Environmental Sciences, Ca' Foscari University,
Dorsoduro 2137, Venezia - 30123, Italy)
ABSTRACT
During
a research on the dynamics and environmental quality of the water-sediment
compartments in the Venice canal network, a study on heavy metal and suspended
particulate matter (SPM) was performed in a pilot subsystem. Measurements of
current velocity were linked to a study on heavy metals in the SPM collected by
large volume filtration and sediment traps. Total heavy metal concentrations
and grain-size distribution were determined on materials yielded by the two
procedures, providing useful information on SPM dynamics in the water column
and an estimate of contaminant fluxes. The study revealed as only a minor part
of the polluted particulate vehiculated in the network, and continuously
resuspended by the effect of tidal currents and boat traffic, undergoes to
deposition within the system.
INTRODUCTION
The City of Venice (Northern Italy), which appears in Figure 1, is
located in the central part of a coastal lagoon, a 550 km2 body of
water connected to the Adriatic Sea by three inlets. The city is traversed by
an intricate canal network, with a linear extension of about 40 km and a mean
depth of 1 m, which is characterised by a complex water circulation pattern
determined by the tide. The average tide excursion ranges from about 20 to 80
cm, respectively in neap and spring tide conditions and salinity generally
ranges from 27 to 33 psu. Water exchanges are mainly driven by the Canal
Grande, which is connected to the sea by first order tidal channels originating
from the Lido inlet.
A large amount of organic
matter and pollutants from domestic effluents and commercial activities are
discharged in the canal network, in which the slack hydrodynamics favours the
progressive accumulation of reduced (Eh » -200 mV) and highly contaminated sludge. Channel silting determine
undesirable impacts on navigation and hygienic conditions.
The
maintenance operations for the canal network include dredging and the
improvement of the sewer system. However, due to the complex morphology and the
short term variability induced by tide, it is quite difficult to define
reliable criteria to assess the effects of the maintenance operations on the
water quality. A preliminary understanding of the mechanisms of transport, the
deposition of particulate matter, and the associated pollutants, is therefore
necessary to investigate the input and fate of the contaminants in this system.
With this aim, a study on heavy metals in the suspended particulate matter
(SPM) was performed in a pilot subsystem (Figure 1) consisting of three
interlinked canals fed by the Canal Grande. The characteristics of suspended
particulate matter, as well as metal concentrations, were investigated by
different techniques including small and large volume filtration (SVF, LVF) and
sediment traps. The research activity was conducted on behalf of the Consorzio
Venezia Nuova – Magistrato alle Acque di Venezia (Italian Ministry of the
Public Works).
METHODS
The
study was conducted during a 36-hours period from the 18th through
the 19th of October 1997, during a particularly high spring tide
excursion (about 100 cm). Continuous measurement of current velocity was
performed in Section A (Figure 1) by a self recording current meter positioned
at a depth of 50 cm near the bottom of the canal.
The
trend of SPM was investigated by small volume filtration (SVF) performed at
hourly intervals following a half –tide cycle from 7.00 to 19.00 on October 19th.
Water samples from surface and bottom layers were filtered on pre-weighed 0.4
µm polycarbonate membranes and the concentration of SPM was obtained by weight
loss after drying the samples at the temperature of 105 °C.
To
obtain representative samples of the SPM in sufficient amounts to permit heavy
metal and grain-size analysis, five LVF runs were executed at almost regular
intervals in the same period covered by SVF. A volume of approximately 2 m3
of canal water, collected from the half depth of the water column, was filtered
through 0.5 µm cartridges during 1 hour interval for each LVF run. The material
was successively recovered by ultrasonic washing and centrifugation.
Sediment traps with a 2÷1
H/W geometry (Gardner 1980 a, b) were placed at an height of about 30 cm from
the canal bottom in two sections (Sites 1 and 2, Figure 1) and left in place
for an overall period of 15 days. Two sampling periods (6 and 9 days) were
established in order to differentiate the fluxes of materials in neap (first
period) and spring (second period) tide conditions.
The
particulate yielded by LVF and sediment traps was analysed for total heavy
metal concentrations (Cd, Cu, Fe, Mn, Ni, Pb, Zn) by A.A.S. after a 8M HN03
microwave digestion; grain-size distribution was also determined by a laser
particle-size analyser.
RESULTS AND DISCUSSION
The trend of the axial component of the current velocity and tide
level acquired by the current meter in Section A are shown in Figure 2.
Velocity data has been obtained from a smoothing procedure of the first order,
to reduce the fluctuations present in
the original records. In the early flood, the flow is directed northward and
the tide forcing from the southern bend of Canal Grande transports water from
the inner part of the canal network. The SPM concentrations, reported in Figure
3, increases particularly in the lower layer, in which resuspension events
occur in the correspondence of the maximum current velocity. In the late flood,
clearer water coming from the southern bend of Canal Grande, which is
relatively closer to the Lido sea inlet, pervades the system. The observed
decrease of SPM concentration and the disappearance of vertical gradients is a
consequence of both dilution and the inhibition of resuspension phenomena by
the high water head and the slack dynamics. During the ebb phase, the system is
drained from the south drawing water from the northern bend of Canal Grande.
The progressive increase of current speed induces an increase of SPM concentrations
and the vertical gradient is again observed. A peak of concentrations with a
uniform vertical distribution is measured in the minimum tide phase (about 22
mg/l around 19.15) as a consequence of the resuspension in the low water head.
The trend of currents and SPM data evidence the role of resuspension by shear
stress and boat traffic, in controlling the presence of particulate in the system.
Heavy metal concentrations
in the particulate collected by LVF and sediment traps are reported in Table 1.
Although LVF samples were collected following a high amplitude tide cycle, the
values display a quite constant trend. Metal concentrations are generally high;
in particular the levels of Zn and Cu is the result of the continuous input of
contaminants from different sources (Zaggia and Zonta, 1997).
The
materials collected by sediment traps in the two different sampling periods
show an even more constant trend of concentrations. Standard deviations for LVF
and sediment traps samples is respectively of the order of 10% and 5% of the
average values.
|
|
|
|
|
|
|
|
|
|
|
|
|
Fe |
Mn |
Cd |
Zn |
Cu |
Pb |
Ni |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
LVF |
Run 1 |
29200 |
411 |
3.5 |
732 |
193 |
128 |
29.8 |
|
|
Run 2 |
25200 |
434 |
2.4 |
699 |
181 |
131 |
25.9 |
|
|
Run 3 |
30700 |
458 |
2.4 |
602 |
174 |
127 |
26.0 |
|
|
Run 4 |
31800 |
456 |
2.4 |
659 |
154 |
134 |
24.4 |
|
|
Run 5 |
25300 |
355 |
2.5 |
600 |
151 |
127 |
20.1 |
|
|
Average |
28440 |
423 |
2.6 |
658 |
171 |
129 |
25.2 |
|
|
st.dev |
3055 |
42 |
0.5 |
58 |
18 |
3 |
3.5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Trap 1 |
Period 1 |
29700 |
408 |
2.6 |
802 |
192 |
156 |
20.3 |
|
|
Period 2 |
29200 |
379 |
2.9 |
1060 |
203 |
166 |
20.6 |
Trap 2 |
Period 1 |
28800 |
390 |
2.8 |
962 |
188 |
167 |
19.4 |
|
|
Period 2 |
27400 |
361 |
2.8 |
988 |
190 |
151 |
19.6 |
|
|
Average |
28775 |
385 |
2.8 |
953 |
193 |
160 |
20.0 |
|
|
st.dev |
988 |
20 |
0.1 |
109 |
7 |
8 |
0.6 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Canal sludge |
Average |
25338 |
367 |
3.3 |
732 |
195 |
162 |
24.2 |
|
“s” Sites |
st.dev |
4215 |
135 |
1.8 |
277 |
64 |
70 |
14.8 |
|
|
|
|
|
|
|
|
|
|
Table 1. Heavy metals concentrations (mg/kg, d.w.) in the particulate collected by LVF and sediment traps; for comparison, average data from the canal sludge of the area are reported .
Table
1 also shows, for comparison, the average concentrations found in the sludge
from canals of the inner part of the investigated system. Data refers to 40
cm-long sediment cores collected in the sites indicated with “s” in Figure 1
(from the Venice Municipal Authority’s database). Despite the close correspondence
of average values among the sludge samples and SPM obtained by both LVF and
sediment traps, a great variability characterises the metal distribution in the
sludge of the canal bottom. This variability is the direct consequence of the
presence of a large number of contaminant sources of and the poor vertical
homogeneity of the collected cores.
As
for heavy metals, the grain-size spectra of
the two groups of 5 LVF and 4 traps samples show no significant
differences; therefore each of the averaged distributions is representative of
the relative set of samples (Figure 4). Besides a partial overlap in the finest
region and the occurrence of two peaks
centred on 1.5 and 7 µm in both the distributions, materials collected by the
traps show a coarser spectrum with a main peak around 30 µm and about 40 % of
the particles in the range of diameter comprised between 20 and 300 µm. This
feature is an expected consequence of the greater efficiency of sediment traps
in collecting coarser materials entrained by turbulence in the near-bottom
layer.
The
flux of particulate recorded by sediment traps is higher in spring tide (130
and 170 gm-2d-1 for trap 1 and 2) than in neap tide
conditions (respectively 97 and 123 gm-2d-1). In spring
tide conditions current speeds are, in fact, more intense and the water head
frequently reaches very low levels favouring entrainment and transport in the
system. The higher yield obtained in Site 2 is instead primarily related to
resuspension induced by boat manoeuvring in a narrowing of the canal section.
Sediment traps data can be finally used for a comparison of the fluxes
of suspended particulate matter with the average sedimentation rates. Typical
values estimated for the canal network from the rate of silting and
radio-dating techniques, are of the order of 1 cmy-1. The fluxes
obtained by sediment traps return instead a value about 3 times greater,
suggesting that only a minor amount of materials transported in the system
constitutes the net sedimentation. Therefore, the major part of SPM undergoes a
frequent reworking by both natural and man-induced resuspension.
The continual cycling of highly contaminated particles in the
water column of a partially confined environment, such as the investigated
subsystem, represents a risk for the long-term effects associated with the
release of toxic elements (Collavini et al., 2000). These particular aspects
have been already considered in the study of analogous systems and will be
further investigated by detailed researches.
REFERENCES
Collavini F, Zonta R, Arizzi Novelli A and
Zaggia L (2000), Environ. Toxicol. Chem., Submitted.
Gardner
WD (1980a), J. Mar. Res. 38:17-39.
Gardner
WD (1980b), J. Mar. Res. 38:41-52.
Zaggia
L and Zonta R (1997), Applied Geochem. 12:527-536.