A recent history of metal accumulation in the sediments of Rijeka harbor, Adriatic Sea, Croatia

Page 1

Marine Pollution Bulletin 62 (2011) 154–167

Contents lists available at ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier .com/locate /marpolbul

A recent history of metal accumulation in the sediments of Rijeka harbor,Adriatic Sea, Croatia

Neven Cukrov a,⇑, Stanislav Franciškovic-Bilinski a, Bojan Hlaca b, Delko Barišic a

a Ru -der Boškovic Institute, Bijenicka 54, Zagreb, Croatiab Port of Rijeka Authority, Riva 1, Rijeka, Croatia

a r t i c l e i n f o

Keywords:Rijeka harbor (Croatia)MetalsPort sedimentsSediment quality

0025-326X/$ - see front matter � 2010 Elsevier Ltd.doi:10.1016/j.marpolbul.2010.08.020

⇑ Corresponding author. Tel.: +385 98 708174; fax:E-mail addresses: [email protected] (N. Cukrov), fra

Bilinski), [email protected] (B. Hlaca), d

a b s t r a c t

We studied metal pollution in the sediments of Rijeka harbor, including anthropogenic influence duringrecent decades and at the present time. Sediment profiles were collected at ten sampling points. The con-centrations of 63 elements in bulk sediment were obtained using ICP–MS, and the concentrations ofselected elements were evaluated by statistical factor analyses. We also calculated metal-enrichment fac-tors and geoaccumulation indices and constructed spatial-distribution maps.

Mercury (Hg) was the heaviest pollutant, with concentrations exceeding 4 mg/kg. Silver (Ag) was thesecond most important pollutant, with constantly increasing values. The average concentrations of themost toxic elements were comparable to those found in sediments of other ports throughout the world,and their toxicity ranged from threshold values [chromium (Cr), arsenic (As)] and midrange-effect values[cadmium (Cd), lead (Pb), copper (Cu), zinc (Zn), nickel (Ni)] to extreme-effect values (Hg). Metal pollu-tion has decreased during recent decades, except for Ag and barium (Ba).

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Metals introduced by human activities into the marine environ-ment accumulate in sediments and are therefore useful indicatorsof anthropogenic inputs. For this reason, several studies aimed atevaluating heavy-metal contamination in harbors have focusedon the sediment compartment (Covelli et al., 2006; Denton et al.,2005; Guerra-Garcia and Garcia-Gomez, 2005; Huerta-Diaz et al.,2008; Sprovieri et al., 2007; Zonta et al., 2007). A fundamentalcharacteristic of trace metals is their lack of biodegradability. Onceintroduced into the aquatic environment, trace metals are redis-tributed throughout the water column, deposited or accumulatedin sediments and consumed by biota. However, sediments arenot only a sink but also a possible delayed source of these contam-inants into the aquatic phase due to desorption and remobilizationprocesses (Fichet et al., 1998; Long et al., 1996). Clearly, sedimentsconstitute a long-term source of contamination to the food web(Burton, 2002).

To evaluate trace-metal pollution in sediments, it is importantto distinguish the natural presence of trace elements due to sedi-mentation from the impact of anthropogenic activities.

In this study, we examined the depth profiles of trace metals incore sediments collected from Rijeka harbor, Croatia, to evaluate

All rights reserved.

+385 1 [email protected] (S. Franciškovic-

[email protected] (D. Barišic).

the anthropogenic inputs over recent decades. To assess the poten-tial sediment toxicity due to metal pollution in the study area, wecompared bulk-sediment concentrations to published sedimentcriteria and standards (Burton, 2002). We also calculated metal-enrichment factors (EF) and geoaccumulation indices.

Our goal was to investigate a possible correlation betweenindustrial activities in the past and the environmental quality ofthe Rijeka harbor area and to assess the levels and types of contam-ination in relation to the geographical distribution and primarysources of pollution. The results of this study will serve to informdecision-makers about the sources and distribution of sediment-associated contaminants and their potential environmental im-pacts in the region and to assess the potential sediment toxicitydue to metal pollution. The results will be applicable to science-based policy in integrated coastal management (ICM) and in thedevelopment of Croatian sediment-quality guidelines.

2. The study area

The Port of Rijeka is located in the northern part of the easterncoast of the Adriatic Sea, in Kvarner Bay. The port can be ap-proached through the Large, Middle and Small gates, which are lo-cated between the Istrian peninsula and the island of Cres,between the islands of Cres and Krk and between the island ofKrk and the shore, respectively. The first dated record of historicaloperations in Rijeka harbor was in 1281. Since then, the harbor andport operations have grown so that Rijeka harbor is now the

Page 2

N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167 155

primary and largest port in Croatia (Dubravic, 2001). The port areais spread over the following five basins: Rijeka, Sušak, Bakar, Rašaand Omišalj. In this paper, we focus on the Rijeka and Sušak basins.

Since the Second World War, the harbor area has been sub-jected to environmental pollution from increasing urban andindustrial development. Possible major sources of pollution inthe area of Rijeka harbor (including the Rijeka and Sušak basins) in-clude municipal wastewaters, port facilities, an oil refinery and apaper mill (Pravdic, 1995). The City of Rijeka uses combined sew-erage systems for wastewater discharge. In these systems, domes-tic, commercial and industrial effluents and some rainwater arecombined before being discharged into the sea. These combinedsewerage systems may contribute significantly to the total concen-trations of a wide range of metals (Chon et al., 2010). However, in-puts of contaminants to Rijeka harbor have been greatly reduced inrecent decades due to the improvement of wastewater treatmentfacilities, removing inflows from the harbor area; the closure ofthe paper mill and modernization of the harbor facilities; and in-creased public awareness.

3. Materials and methods

3.1. Sampling

Sediment samples were collected at ten sampling points in June2008 using Uwitec gravitational corers (u = 6 cm) equipped withPVC tubes (120 cm). Depending on the bottom conditions, sedi-ment columns up to 35 cm were sampled. Sampling density wasadapted to the characteristics of Rijeka harbor (Fig. 1). A total of10 sampling locations (nine inside the harbor area and one located2 km outside the harbor) were mapped using a GPS receiver (Gar-min GPS Map 76 CSx; Kansas City, USA) with an accuracy of ±5 m.A map of the study area showing the sampling points is provided inFig. 1.

Sediment columns were cut into 5-cm slices and kept in plasticbags at 3 �C until analysis. This sediment-slice thickness was foundto be optimal to determine relevant sediment dates related to par-ticular sedimentation conditions.

3.2. Sediment analysis

For granulometric analysis, half of each sediment sample waswet sieved using ambient water in AS 200 Digit – RETSCH sieveshakers with analytical sieves (63, 125, and 250 lm; 1, 2, and4 mm). The other half of each bulk-sediment sample was dried

Fig. 1. Map of the

and separated for future analysis. The Folk (1954) grain-size classi-fication was used.

Chemical analyses were performed at ACTLABS (www.actl-abs.com), Ontario, Canada, using ICP, ICP/MS and cold-vapor AAfor mercury (Hg) determination. The sample material was digestedwith aqua regia (0.5 ml H2O, 0.6 ml concentrated HNO3 and 1.8 mlconcentrated HCl).

The concentrations of 63 elements were measured. Those of Hg,lithium (Li), beryllium (Be), aluminum (Al), phosphorus (P), chro-mium (Cr), nickel (Ni), copper (Cu), zinc (Zn), arsenic (As), scan-dium (Sc), silver (Ag), cadmium (Cd), tin (Sn), antimony (Sb),lanthanum (La), cerium (Ce), europium (Eu), lead (Pb) and uranium(U) were evaluated by statistical factor analyses.

Prior to gamma-spectrometry measurements, selected bulk-sediment samples (about 150 g) were thawed at room tempera-ture, dried at 106 �C for 24 h (to constant mass), and placed in ahome-made counting vessel (polyethylene, 125 cm3). The vesselwas sealed and stored for at least 4 weeks to allow radiochemicalequilibrium between 226Ra and its gaseous 222Rn progeny.

The 137Cs activity of each sample was determined by gammaspectrometry using a low-background hyper-pure germanium(HPGe) semiconductor detector system (‘‘Canberra”) coupled toan 8196-channel analyzer (Meriden, USA). Characteristics of thesystem were as follows: FWHM for 1.33 MeV (60Co) = 1.76 keV;relative efficiency for 1.33 MeV = 25.4%; photo-peak to Comptonratio for 1.33 MeV = 52.3:1. The system was calibrated using cali-bration standards (stream sediments 313 and 314) supplied bythe IAEA. All results are presented with 95% confidence intervals.Spectra were usually measured for 80,000 s and were analyzedby the Canberra GENIE 2K software. The 137Cs activities were calcu-lated over a 661.6-keV peak.

Sedimentation rates were estimated using the vertical distribu-tion of 137Cs activity in sediment columns. In the study area, twoperiods of huge 137Cs input into the environment have occurred.The first was caused by French atmospheric nuclear testing inthe Sahara in the early 1960s; the second was caused by the nucle-ar accident at Chernobyl in April 1986. We used the positions ofthese two 137Cs-activity peaks in the sediment columns to estimatesedimentation rates (Cukrov et al., 2009; Sanders et al., 2006; Weiet al., 2007).

3.3. Environmental relevance of sediment metal concentrations

Multiple approaches were used to assess the potential sedimenttoxicity derived from metal pollutants in the study area:

sampling area.

Page 3

156 N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167

(1) Metal-enrichment factors (EF).(2) Geoaccumulation indices (Igeo).(3) Comparisons of bulk-sediment values to published sediment

criteria and standards.

We also evaluated the distribution of element concentrations inthe sediment samples by statistical analyses, including calculationof basic statistical parameters and cluster, boxplot and factoranalyses).

3.3.1. Enrichment factorsGeochemical normalization based on the concentration of a

conservative element such as iron (Fe) is commonly used to iden-tify anomalous metal concentrations. The enrichment factor (EF) isdefined as the observed metal/Fe ratio in the sample of interest di-vided by the background metal/Fe ratio (Guerra-Garcia and Garcia-Gomez, 2005). Because Fe is one of the most abundant elements onearth and usually poses no contamination concern, it is the mostfrequent choice for normalization purposes. The enrichment factor(EF) is expressed as

EF ¼ ðMe=FeÞsampleðMe=FeÞbackground

where (Me/Fe)sample is the metal/Fe ratio in the sample of interestand (Me/Fe)background is the natural background value of the me-tal/Fe ratio. For the natural background value, we chose the deepestsediment layer from the reference sampling site (LR7). This sedi-ment was deposited 100 years ago, which was during the preindus-trial era in that zone. We calculated EF values for Hg, Cd, Pb, Cu, Zn,Cr, Ni, Co, As, Ag, Mo, Sn and Ba.

Five degrees of contamination are commonly defined (Suther-land, 2000):

EF < 2, deficiency to low enrichment;EF 2–5, moderate enrichment;EF 5–20, significant enrichment;EF 20–40, very high enrichment; andEF >40, extremely high enrichment.

Enrichment-factor contour maps were constructed to illustratespatial variations in the datasets using for Windows (Golden Soft-ware Inc., version 8). The Kriging algorithm was applied as theinterpolation method. Maps were constructed based on EF calcula-tions (for Hg, Cd, Pb, Cu, Zn, Cr, Ni, Co, As, Ag, Sn and Ba) in recentsediments and in 25-year-old sediment layers. The latter periodwas the oldest possible from our sediment columns and representsthe time period during which the strongest anthropogenic influ-ence probably occurred.

3.3.2. Geoaccumulation indexA companion method for identifying concentrations of concern

is the geoaccumulation index (Igeo), which was originally intro-duced by Müller (1979):

Igeo ¼ log2ðMnÞ

1:5ðBnÞwhere Mn is the measured concentration of the examined metal (n)in the sediment and Bn is the geochemical background concentra-tion of the metal (n). A factor of 1.5 is used as the background ma-trix-correction factor due to lithogenic effects. The backgroundvalues of the metals of interest are the same as those used in theenrichment-factor calculation described above. Müller (1979) hasdefined seven classes of the geoaccumulation index, ranging fromClass 0 (Igeo 6 0, unpolluted) to Class 6 (Igeo > 5, extremely pol-luted). The highest class (Class 6) reflects at least 100-fold enrich-ment above background values.

3.3.3. Sediment-quality criteriaNumerous sediment-quality guidelines have been developed

during the past 20 years to assist regulators in dealing with con-taminated sediments. Unfortunately, most of these guidelines havebeen developed in North America and have errors of 25% or more(Burton, 2002). Because of the lack of Croatian regulations, wecompared metal concentrations in our sediment samples to vari-ous marine-sediment-quality guidelines (Burton, 2002).

3.4. Statistical analyses

All statistical analyses were performed using Statistica 6.0 (Stat-Soft, 2001, http://www.statsoft.com). The following statisticalanalysis was performed:

(a) To summarize the experimental data, we calculated basicstatistical parameters including N (number of cases), mean,geometric mean, median, mode, frequency, minimum, max-imum, standard deviation, skewness, and kurtosis.

(b) The boxplot method was used to identify anomalies in sed-iment samples. Normal or lognormal boxplots were con-structed from the empirical cumulative–distribution plots.The box length represented the interquartile range, whileoutlier values were defined as those lying between 1.5 and3 box lengths from the upper or lower edge of the box.Extreme values were those more than three box lengthsfrom the edge of the box (Reimann et al., 2005; Tukey, 1977).

(c) A cluster analysis of Q-modality was performed to findgroups containing similar samples. A cluster analysisbelongs to a multivariate statistics and represents a hierar-chical method. There are two modes of cluster analysis: Q-modality, in which clusters of samples are sought, and R-modality, in which clusters of variables (in our case, ele-ments) are obtained. Hierarchical agglomerative clusteranalysis begins by calculating a matrix of distances amongitems in the data matrix. In Q-modality analysis, the distancematrix is a square that expresses all possible pairwise dis-tances among samples. A variety of distance metrics canbe used to calculate similarity. For data that show linearrelationships, the Euclidean distance is a useful measure.More details about cluster analysis can be found in Kaufmanand Rousseeuw (1990). Our samples were grouped intothree clusters, and clustering was performed using all chem-ical elements from the database.

(d) Factor analysis was also performed. This method differs fromPCA in that it is usually considered a statistical technique.From ‘‘regular observations,” the PCA method extracts themain elemental associations that control the multivariateassignment of the database, while the factor-analysis proce-dure assumes that the relationship between a set of m vari-ables reflects the correlation of each variable with pmutually non-correlated main factors. The general assump-tion is that p < m. The variance of m variables is derived fromthe variance of the p factor. More details about factor analy-sis can be found in Davis (1986) and Halamic et al. (2001).

4. Results and discussion

4.1. Grain size

Sediment grain-size fractions were classified as fine (silt–clay;<0.063 mm), sand (0.063–2 mm) and gravel (>2 mm). Average per-centages of gravel, sand, and fine particles and median grain sizefor each sediment column are presented in Table 1. The fine frac-tion was generally predominant (>50%) in the sediment columns.The spatial distribution of grain sizes shows that coarser sediments

Page 4

were

locatedin

front

ofth

eR

ijekaoil

refin

ery(LR

5)an

dn

earth

em

outh

ofth

eR

jecina

River

(LR9).

4.2.Sedimentation

rate

Sedimen

tationrates

were

estimated

usin

gth

everticaldistribu

-tion

of1

37C

sactivity

(Cu

krovet

al.,2009;San

derset

al.,2006;W

eiet

al.,2007)in

the

sedimen

tcolu

mn

s(Fig

2).The

estimated

ann

ual

sedimen

tationrates

were

5–6m

mfor

the

principal

harbor

area(LR

1–LR6),6–7

mm

inth

esecon

daryarea

(LR8–LR

10)an

d3–4

mm

atthe

reference

point2

kmfrom

the

harbor(LR

7).Usin

gth

esecalcu

-lated

sedimen

tratesto

calibrateth

eavailable

data,we

separatedou

tth

esedim

ent

layersrepresen

ting

the

periodfrom

1983to

1985.

4.3.Toxicm

etals

4.3.1.Enrichment

factor(EF)

Table2

presents

the

ranges

ofEF

values

for13

selectedelem

ents.

Table1

Average

valuesofgravel,sand,fi

neparticles

andm

ediangrain

sizefor

sediment

coresof

theRijeka

harbor.

Nu

mber

ofsu

bsamples

Gravel

(%)

Sand

(%)

Fine

(%)

Median

(mm

)

LR1

75.64

33.3261.04

0.042LR

26

3.9439.26

56.790.058

LR3

77.25

34.7757.98

0.051LR

46

6.2127.53

66.270.037

LR5

314.39

49.4236.19

0.24LR

65

8.125.15

66.750.03

LR7

716.52

16.8566.63

0.046LR

87

7.2539.23

53.520.047

LR9

44.47

27.1768.37

0.042LR

104

4.4743.58

51.940.125

Fig.2.V

erticaldistribu

tionof

13

7Cs

activity(B

qkgdry

weigh

t).

Cr Ni Co As Ag Mo Sn Ba

SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD

1.94 1.10 0.07 1.16 0.08 0.93 0.09 1.47 0.12 28.4 8.58 3.07 0.45 5.62 0.86 1.65 0.2210.5 1.00 1.23 1.03 1.28 0.80 1.07 1.29 1.67 16.4 36.9 2.35 3.54 4.34 6.51 1.36 1.941.57 1.08 0.06 1.09 0.04 0.89 0.07 1.70 0.32 25.9 12.4 5.99 2.17 6.77 1.25 2.03 0.4710.4 1.02 1.16 1.03 1.15 0.83 0.98 1.31 2.13 13.7 43.2 3.60 9.41 5.42 8.81 1.60 2.922.87 1.08 0.06 1.10 0.07 0.90 0.09 1.55 0.30 24.4 12.6 3.67 1.36 4.97 0.76 2.13 0.4710.5 0.97 1.16 1.03 1.20 0.82 1.04 0.92 1.91 5.74 37.9 1.56 5.48 3.85 5.70 1.61 2.784.25 1.10 0.03 0.97 0.02 0.79 0.03 1.84 0.18 15.7 2.03 4.60 2.30 6.89 2.81 2.66 0.5418.50 1.07 1.16 0.94 0.99 0.74 0.84 1.66 2.17 12.0 18.0 2.09 7.78 4.03 11.1 2.11 3.60

N.Cukrov

etal./M

arinePollution

Bulletin62

(2011)154–

167

�1

Table 2Ranges of enrichment factor for 13 selected elements.

Sample Hg Cd Pb Cu Zn

Mean SD Mean SD Mean SD Mean SD Mean

LR1 139 22.8 11.6 1.80 10.7 1.13 8.41 3.67 6.27Range 108 168 9.78 14.5 9.31 12.7 5.24 16.2 5.02LR2 100 18.8 20.7 5.33 13.2 2.57 7.07 1.54 8.31Range 80.5 132 11.3 27.0 10.2 17.3 5.23 9.17 5.67LR3 83.3 25.9 16.9 8.73 13.4 5.83 6.32 1.44 6.95Range 41.5 120 3.41 26.9 4.56 23.2 3.60 7.54 2.69LR4 37.6 4.05 24.4 13.5 20.7 5.51 7.06 0.92 11.0Range 32.7 43.6 13.3 48.9 15.6 29.4 6.00 8.11 7.70

LR5 28.7 1.53 7.10 4.52 24.3 0.83 7.01 0.92 7.57 0.74 0.98 0.02 0.78 0.05 0.70 0.04 2.51 0.11 8.49 0.88 2.74 0.62 7.51 4.55 3.97 0.29Range 27.0 29.9 4.43 12.3 23.5 25.2 6.32 8.05 6.73 8.10 0.95 1.00 0.73 0.83 0.66 0.73 2.39 2.59 7.91 9.51 2.36 3.46 3.76 12.6 3.64 4.15LR6 25.8 7.18 4.10 0.62 11.0 3.71 3.97 0.49 3.06 0.49 1.02 0.06 1.11 0.01 0.91 0.02 1.17 0.10 9.78 4.09 1.84 0.39 4.09 1.31 3.09 0.72Range 16.1 35.9 3.33 4.66 7.86 17.0 3.20 4.41 2.21 3.41 0.93 1.09 1.09 1.12 0.89 0.93 1.02 1.27 4.65 14.4 1.23 2.19 2.30 6.00 2.30 3.81LR7 5.09 2.60 1.22 0.14 1.65 0.45 1.42 0.23 1.36 0.22 1.00 0.01 1.05 0.04 1.02 0.02 0.76 0.13 1.05 0.25 1.00 0.06 1.29 0.21 1.16 0.16Range 1.00 8.10 1.00 1.43 1.00 2.27 1.00 1.70 1.00 1.67 1.00 1.02 1.00 1.11 0.99 1.05 0.63 1.00 0.63 1.40 0.90 1.08 1.00 1.54 0.99 1.42LR8 154 53.4 14.3 2.85 11.9 2.03 10.1 0.95 7.11 1.34 1.37 0.07 1.39 0.07 0.93 0.06 1.48 0.18 47.3 11.8 4.40 1.58 6.01 0.50 1.32 0.24Range 96.7 245 11.0 18.5 10.3 15.4 9.00 11.7 5.95 9.42 1.28 1.50 1.32 1.51 0.88 1.03 1.34 1.74 30.6 61.2 2.69 6.95 5.34 6.58 0.96 1.58LR9 12.4 2.42 9.98 2.32 3.70 0.53 3.42 0.30 2.18 0.21 1.16 0.03 1.07 0.06 0.82 0.04 1.11 0.05 7.38 2.89 3.44 0.18 2.72 0.81 1.96 0.24Range 8.84 14.3 7.90 12.3 2.98 4.24 3.06 3.67 1.88 2.36 1.12 1.19 1.00 1.13 0.77 0.86 1.06 1.19 4.79 9.98 3.19 3.59 2.22 3.93 1.64 2.19LR10 7.34 2.71 2.32 0.45 1.97 0.58 1.86 0.25 1.51 0.28 1.03 0.04 1.18 0.01 1.04 0.02 0.91 0.10 2.85 1.45 1.46 0.42 2.24 0.62 1.43 0.32Range 3.35 9.38 1.77 2.72 1.31 2.57 1.57 2.10 1.24 1.77 1.00 1.08 1.17 1.20 1.02 1.07 0.83 1.01 1.50 4.28 1.11 1.96 1.36 2.76 1.07 1.73Harbor 76.7 58.4 13.5 8.99 12.4 6.62 6.54 2.86 6.34 3.44 1.12 0.13 1.12 0.16 0.89 0.10 1.52 0.41 21.7 15.7 3.64 1.84 5.33 2.18 2.14 0.82Range 3.35 245 1.77 48.9 1.31 29.4 1.57 16.2 1.24 18.50 0.93 1.50 0.73 1.51 0.66 1.07 0.83 2.59 1.50 61.2 1.11 9.41 1.36 12.6 0.96 4.15

157

Page 5

158 N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167

The spatial distributions of EF values for nine selected metals insurface and historic sediments (sediment layers representing1983–1985) are shown in Fig. 3.

Based on the EF values, Hg is the heaviest pollutant among themeasured elements in the Rijeka harbor area (Table 2). All sedi-

Fig. 3. Spatial distributions of the enrichment factors of nine metals

ments layers of the three sediment columns (LR1–LR3) collectedin the inner-harbor basins exhibited extremely high Hg enrich-ment (>40). The highest values were observed in deeper sedimentlayers of the profile LR8, which is located near an outflow site thatwas used for the municipal-sewage outflow until 1994. Lower EF

in surface and historic sediments representing the last 25 years.

Page 6

N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167 159

values, but still exceeding the value considered to represent extre-mely high enrichment, were found in shallower (more recent) sed-iment layers in this profile, indicating an improving trend sincemunicipal-sewage outflow was abandoned. Most sediment layersof the profiles LR4–LR6, which were collected in front of the oilrefinery, exhibited very high Hg enrichment (20–40). Profiles LR9and LR10, which were collected in front of the container terminalof Rijeka port in the eastern part of the Rjecina River Delta, exhib-ited significant enrichment (5–20). The representative marine-sed-iment sample (LR7) showed significant enrichment in shallower(more recent) sediment layers, while deeper (older) sediment lay-ers exhibited moderate enrichment (2–5). Fig. 3A shows that re-cent sediment layers had much lower EF value than oldersediment layers, indicating decreasing Hg pollution in the Rijekaharbor area. In the historic sediment layer, the center of Hg pollu-tion is clearly visible around profile LR8, located in front of the Rje-cina River Delta and the abandoned municipal-sewage outflowsite. This center of Hg enrichment is not present in the spatial pro-file of the most recent sediments; instead, the highest values arefound in the inner harbor (LR1).

The EF values for Cd in the three inner-harbor profiles (LR1–LR3) show increasing pollution during the past, with valuesexceeding 20, which indicates a very high enrichment in some sed-iment layers (Table 2). Cd pollution has decreased over the last20 years, dropping to values below 20 (significant enrichment). Asimilar situation can be seen in profiles LR4 and LR5, located infront of the oil refinery. In profile LR4, the second-deepest layershows values greater than 40, indicating extremely high enrich-ment. It is interesting that in profile LR6, located only about100 m toward the open sea from the previous two profiles, EF val-ues for Cd are low, within the range of moderate enrichment. Atlocation LR7, which represents marine sediments from this region,all EF values of Cd are less than 2, indicating deficiency to lowenrichment. All sediment layers of profile LR8 show values charac-teristic of significant enrichment (5–20). In the shallowest (mostrecent) layer of profile LR8, there is an obvious trend of decreasingof Cd pollution since the municipal-sewage outflow site locatednear this location was abandoned in 1994. At location LR9, nearthe Rjecina River Delta and in front of the container terminal, thereis a significant trend of decreasing EF values for Cd, but these val-ues remain in the range 5–20, indicating significant enrichment. Inprofile LR10, just a bit farther from the Rjecina River Delta, EF val-ues for Cd are low (slightly above 2), indicating moderate enrich-ment. Fig. 3B shows that the Cd-pollution situation has improvedin recent years. In both sediment layers (recent and historic), thecenter of Cd enrichment is located in the central part of the in-ner-harbor basin (around profile LR2), probably due to port activi-ties. Cd enrichment is significantly lower in recent sediments.

The highest EF values for Pb are found in deeper sediment layersin profiles LR4 and LR5, in front of the oil refinery, and in one of thedeeper sediment layers of profile LR3, which is closest to the refin-ery among the inner-harbor sediment layers (Table 2). These high-est values are in the category of very high enrichment (20–40). Pbprobably originates from the nearby oil refinery, and there is adecreasing trend in the more recent sediment layers because Pbis no longer used in the oil industry. In the other inner-harbor pro-files (LR1 and LR2), in profile LR6 (about 100 m toward the opensea from LR4 and LR5) and in profile LR8 (near the abandoned sew-age outflow), EF values for Pb are in the category of significantenrichment (5–20). The lowest EF values for Pb are found at thelocations of profiles LR9 and LR10, in front of the container termi-nal of Rijeka port and in the marine profile LR7. At these sites, EFvalues for Pb are less than 2, indicating deficiency to low enrich-ment, and 2–5, indicating moderate enrichment. Fig. 3C shows thatthe center of Pb enrichment is located in front of the Rijeka oilrefinery and that the situation is similar in recent and historic sed-

iment layers. The petroleum industry is a major source of Pb in theenvironment (Chon et al., 2010; Councell et al., 2004; Lough et al.,2005).

In all sediment layers of the innermost harbor locations, EF val-ues for Cu indicate significant enrichment (5–20). In profile LR6, infront of the oil refinery but about 100 m farther toward the opensea, EF values drop to 2–5, indicating moderate enrichment. In pro-files LR9 and LR10, in front of the container terminal, values arealso 2–5, corresponding to moderate enrichment, even droppingto less than 2 (deficiency to low enrichment) in some deeper sed-iments layers of LR10. The marine-sediment sample LR7 has verylow values (<2) in all sediment layers, corresponding to deficiencyto low enrichment. Fig. 3D clearly shows that Cu pollution has de-creased significantly during the last 25 years. The figure showinghistoric sediment layers shows an intensive Cu-enrichment centerin the inner part of the Rijeka port, around profile LR1, but this cen-ter vanishes in the recent sediment layers.

EF values for Zn show similar behavior to those of Cu (Table 2).In most of the harbor profiles, Zn EF values are in the significant-enrichment range (5–20). These profiles show an interesting trendof increasing Zn pollution during the historical phase, followed by atrend of decreasing pollution in the last fifteen to 20 years. As withCu, EF values for Zn in profile LR6 (in front of the oil refinery) and inmost sediments layers in profile LR9 are in the range 2–5, corre-sponding to moderate enrichment, while in profile LR10 these val-ues are less than 2, indicating deficiency to low enrichment, thesame as in the marine profile LR7. Fig. 3E depicts the spatial patterof Zn enrichment. About 25 years ago, a major center of Zn enrich-ment stretched along the whole inner basin of Rijeka port and to-ward the oil refinery. In addition to municipal wastewaters, cityrunoff waters with elevated concentrations of Zn and Cu flowedinto the harbor in that area (Chon et al., 2010; Councell et al.,2004; Thevenot et al., 2007). In recent sediment layers, this centerdecreased in both area and intensity, indicating that Zn pollutionhas also decreased considerably in recent times.

EF values for As were very low (<2) in most locations (Table 2).Values slightly greater than 2 were observed in all sediment layersof profile LR5 (in front of the oil refinery), in one sediment layer ofprofile LR4 (also in front of the oil refinery) and in one layer of pro-file LR2 (from the inner-harbor basin), indicating moderate enrich-ment with As. Fig. 3F shows the spatial distribution of As-enrichment values in historic and recent sediment layers. Recently,As enrichment has decreased slightly, but the spatial distributionhas remained the same, indicating the same source. The center ofAs pollution is positioned in front of the oil refinery; petroleum-refining operations are a known source of As in the environment(Chon et al., 2010). In the historic sediment layers, the As-enrich-ment center also stretched along the inner basin of Rijeka port,while in recent sediment layers, it remained only in front of therefinery.

Ag exhibited the second highest EF values and thus representsthe second-heaviest pollutant in Rijeka harbor (Table 2). Most sed-iment layers of the inner-harbor profiles (LR1–LR3) exhibited veryhigh enrichment factors. However, deeper sediment layers showedmuch lower values, corresponding to significant enrichment. Thispattern indicates increased Ag pollution in the Rijeka harbor basinduring recent decades. In the three profiles in front of the oil refin-ery (LR4–LR6), almost all sediment layers show significant Agenrichment (5–20). The highest EF values for Ag were observedin profile LR8, near the abundant sewage outflow. There is no signof improvement; Ag EF values have increased constantly duringthe recent past, from values indicating very high enrichment inthe deepest sediment layers to extremely high enrichment in new-er sediment layers. A trend of increasing Ag enrichment is alsoshown by profiles LR9 and LR10, located in front of the Rijeka portcontainer terminal. In profile LR9, which is located closer to the

Page 7

Tabl

e3

Calc

ulat

edva

lues

ofIg

eofo

r13

sele

cted

elem

ents

.

Sam

ple

Hg

Cd

PbC

uZn

Cr

Ni

Co

As

Ag

Mo

SnB

a

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

LR1

4.76

0.19

2.28

0.17

2.20

0.11

1.91

0.40

1.65

0.29

�0.

060.

06�

0.02

0.06

�0.

240.

090.

220.

113.

140.

360.

950.

151.

560.

160.

330.

11R

ange

4.52

4.95

2.11

2.60

2.09

2.40

1.49

2.71

1.44

2.28

�0.

170.

01�

0.08

0.06

�0.

35�

0.12

0.09

0.44

2.63

3.53

0.74

1.11

1.34

1.72

0.15

0.48

LR2

4.38

0.18

2.78

0.31

2.35

0.18

1.72

0.21

1.89

0.21

�0.

140.

04�

0.13

0.07

�0.

330.

100.

300.

172.

940.

471.

520.

361.

680.

180.

480.

21R

ange

4.17

4.68

2.19

3.09

2.09

2.58

1.42

1.94

1.50

2.14

�0.

20�

0.07

�0.

24�

0.04

�0.

46�

0.20

0.04

0.49

2.42

3.54

1.05

2.04

1.42

1.94

0.27

0.84

LR3

4.20

0.34

2.47

0.71

2.32

0.49

1.64

0.24

1.67

0.45

�0.

100.

05�

0.09

0.11

�0.

290.

150.

240.

212.

840.

681.

050.

401.

420.

140.

560.

20R

ange

3.61

4.56

1.11

3.18

1.40

3.03

1.17

1.91

0.88

2.23

�0.

16�

0.05

�0.

220.

07�

0.42

�0.

07�

0.19

0.42

1.63

3.41

0.33

1.59

1.23

1.64

0.29

0.90

LR4

3.51

0.15

2.97

0.53

2.89

0.29

1.84

0.17

2.23

0.39

�0.

020.

06�

0.14

0.05

�0.

340.

090.

490.

122.

630.

131.

310.

561.

750.

440.

850.

17R

ange

3.35

3.71

2.41

3.79

2.60

3.28

1.65

2.01

1.90

2.82

�0.

100.

05�

0.22

�0.

08�

0.47

�0.

240.

380.

682.

422.

750.

561.

951.

222.

360.

701.

14LR

53.

160.

061.

640.

582.

990.

041.

740.

131.

820.

10�

0.22

0.02

�0.

450.

07�

0.56

0.06

0.72

0.04

1.93

0.10

0.79

0.22

1.69

0.61

1.17

0.07

Ran

ge3.

093.

201.

292.

312.

963.

031.

641.

891.

701.

89�

0.25

�0.

20�

0.52

�0.

39�

0.62

�0.

510.

670.

751.

862.

050.

661.

041.

122.

331.

091.

22LR

63.

130.

251.

320.

122.

270.

371.

290.

111.

020.

15�

0.07

0.04

0.01

0.07

�0.

180.

070.

070.

052.

110.

430.

510.

221.

280.

301.

020.

21R

ange

2.75

3.42

1.17

1.48

1.91

2.80

1.14

1.45

0.76

1.13

�0.

10�

0.02

�0.

070.

08�

0.26

�0.

10�

0.01

0.12

1.51

2.53

0.18

0.73

0.80

1.63

0.80

1.29

LR7

1.27

0.82

0.02

0.21

0.29

0.36

0.16

0.27

0.12

0.25

�0.

170.

11�

0.13

0.14

�0.

160.

12�

0.47

0.09

�0.

160.

29�

0.18

0.13

0.06

0.24

�0.

030.

20R

ange

�0.

411.

94�

0.41

0.29

�0.

410.

67�

0.41

0.38

�0.

410.

39�

0.41

�0.

07�

0.41

0.04

�0.

41�

0.02

�0.

61�

0.36

�0.

640.

18�

0.41

0.01

�0.

410.

27�

0.41

0.19

LR8

4.93

0.31

2.59

0.23

2.41

0.15

2.25

0.12

1.89

0.18

0.26

0.10

0.28

0.09

�0.

140.

070.

330.

123.

770.

281.

370.

361.

730.

070.

200.

19R

ange

4.49

5.37

2.27

2.85

2.25

2.67

2.07

2.40

1.67

2.18

0.18

0.43

0.18

0.44

�0.

21�

0.04

0.19

0.49

3.35

4.11

0.91

1.88

1.60

1.81

�0.

100.

38LR

92.

090.

241.

870.

280.

890.

210.

820.

100.

370.

13�

0.26

0.08

�0.

340.

02�

0.61

0.03

�0.

300.

091.

530.

380.

830.

040.

560.

230.

260.

18R

ange

1.74

2.25

1.60

2.15

0.65

1.14

0.68

0.88

0.19

0.48

�0.

32�

0.13

�0.

36�

0.31

�0.

64�

0.57

�0.

41�

0.22

1.12

1.87

0.77

0.86

0.37

0.90

0.05

0.48

LR10

1.50

0.39

0.40

0.10

0.22

0.16

0.19

0.09

�0.

030.

06�

0.39

0.14

�0.

260.

17�

0.38

0.17

�0.

520.

080.

510.

38�

0.08

0.17

0.34

0.22

�0.

090.

10R

ange

0.93

1.83

0.29

0.48

�0.

010.

350.

070.

27�

0.07

0.06

�0.

59�

0.28

�0.

46�

0.11

�0.

58�

0.24

�0.

63�

0.47

0.12

0.86

�0.

170.

170.

020.

54�

0.21

0.02

Har

bor

3.76

1.10

2.17

0.81

2.14

0.80

1.58

0.59

1.49

0.71

�0.

080.

19�

0.08

0.21

�0.

310.

170.

200.

332.

580.

960.

990.

541.

400.

510.

510.

38R

ange

0.93

5.37

0.29

3.79

�0.

013.

280.

072.

71�

0.07

2.82

�0.

590.

43�

0.52

0.44

�0.

64�

0.04

�0.

630.

750.

124.

11�

0.17

2.04

0.02

2.36

�0.

211.

29

160 N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167

Rjecina River Delta, values are higher than in profile LR10, wherevalues are less than 2, indicating deficiency to low enrichment.The marine profile LR7 also exhibits deficiency to low enrichment.The spatial distribution of Ag EF values (Fig. 3G) shows the inversetrend to that of the majority of elements. In the historic layers,there was a small center of Ag enrichment in front of the RjecinaRiver Delta, near the abandoned municipal-sewage outflow. Thesame center appears in the recent layers, but it is much more in-tense exhibits a trend of Ag pollution spreading along the inner-harbor basin of Rijeka port.

Sn also has rather equally distributed EF values in all sedimentlayers of all harbor profiles (Table 2). Values are mostly in therange 5–20, indicating significant enrichment, with some sedimentlayers belonging to the category 2–5, indicating more moderateenrichment. Exceptions are profiles LR9 and LR10, located in frontof the container terminal of Rijeka port and near the Rjecina RiverDelta, where lower values (2–5) prevail, indicating moderateenrichment. In the marine profile LR7, values are less than 2, show-ing deficiency to low enrichment. Fig. 3H clearly shows that about25 years ago, there was a major center of Sn enrichment in the areain front of the Rijeka oil refinery. In the recent sediment layers, thiscenter no longer appears, indicating decreasing Sn pollution.

EF values for Ba are very low in all profiles (Table 2). Values areless than 2 (deficiency to low enrichment) and 2–5 (moderateenrichment). The values are highest in the more recent sedimentlayers of profile LR5 because of the nearby oil refinery, but the EFvalues remain less than 5 and indicate moderate enrichment.Fig. 3I shows the spatial distribution of Ba enrichment factors. Bais one of the few elements whose concentrations have increasedin recent times. The center of Ba enrichment is located in the areain front of the Rijeka oil refinery and occurs in the same position inrecent and older sediments, but the intensity is higher now than inthe past. Ba pollution is known to be connected with diesel fuel(Sutherland, 2000).

EF values for Mo are mostly equally distributed in all harborprofiles (Table 2). Values are mostly in the range 2–5, indicatingmoderate enrichment. However, in some of the deeper sedimentlayers of the inner-harbor stations (LR1–LR3) of profile LR8 (nearthe abandoned sewage outflow) and of sites near the oil refinery,Mo EF values are slightly higher (exceeding 5), indicating signifi-cant enrichment. This pattern shows that Mo enrichment has de-creased recently. In the marine profile LR7, Mo EF values are verylow (less than 2), indicating deficiency to low enrichment.

The elements Cr, Ni and Co show very low EF values (>2) in allsediment layers at all locations, indicating deficiency to lowenrichment (Table 2).

4.3.2. Geoaccumulation indexTable 3 presents the calculated values of Igeo for 13 selected

elements.According to the Igeo values, Hg is the heaviest pollutant among

the measured elements in the Rijeka harbor area. In almost all sed-iment layers of all three profiles taken within inner-harbor basins(LR1–LR3), values of Igeo for Hg are between 4 and 5, indicatingheavily to extremely contaminated sediments. The highest Igeovalues (>5) for Hg, indicating extremely contaminated sediments,are present in deeper sediment layers of profile LR8, located nearthe outflow site used until 1994 for municipal-sewage outflow.In shallower (more recent) sediments layers of this profile, aslightly improving trend can be observed because the sewage out-flow was abundant in 1994 and more recent sediment layers are inthe category between 4 and 5. Notably, the Igeo values indicatethat all other harbor stations are at least moderately or moderatelyto heavily contaminated. Even most sediment layers of profile LR7,chosen as a sample of marine sediments several kilometers fromthe port, have Hg Igeo values between 1 and 2, indicating moder-

Page 8

Tabl

e4

Conc

entr

atio

ns(m

g/kg

)of

sele

cted

met

als

inse

dim

ent

ofRi

jeka

harb

or.

Sam

ple

Hg

Cd

PbC

uZn

Cr

Ni

Co

As

Ag

Mo

SnB

a

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

Mea

nSD

LR1

4.45

0.81

1.04

0.19

217

24.9

207

102

405

147

72.6

4.46

90.9

5.27

13.5

1.22

22.4

2.53

1.57

0.50

3.30

0.46

7.78

1.22

71.5

7.78

Ran

ge3.

435.

300.

871.

4119

326

212

642

931

873

365

.278

.084

.998

.012

.115

.219

.527

.70.

892.

202.

633.

836.

209.

0959

.582

.1LR

23.

040.

571.

760.

4625

444

.916

232

.050

498

.867

.02.

8381

.25.

3712

.41.

1924

.54.

091.

340.

616.

092.

198.

851.

6083

.618

.6R

ange

2.42

4.04

0.94

2.31

193

314

118

199

337

637

63.4

71.6

72.4

88.7

10.8

14.0

18.5

29.0

0.72

2.23

3.59

9.68

6.68

11.3

66.5

118

LR3

2.61

0.79

1.46

0.74

267

120

150

32.6

432

171

69.9

3.13

84.9

9.53

13.0

1.99

23.1

4.12

1.30

0.66

3.85

1.41

6.73

0.95

91.0

19.0

Ran

ge1.

393.

590.

322.

5297

.149

591

.719

218

070

065

.573

.474

.099

.011

.215

.914

.727

.20.

331.

951.

766.

165.

578.

3668

.312

5LR

41.

270.

202.

321.

3444

613

418

131

.274

330

975

.94.

6080

.34.

0812

.21.

0529

.43.

540.

900.

115.

262.

7710

.14.

5812

121

.4R

ange

1.07

1.53

1.17

4.66

321

637

149

213

502

1260

69.6

81.0

73.6

85.2

10.7

13.4

26.0

35.1

0.73

1.01

2.21

8.90

5.49

17.2

103

160

LR5

0.88

0.05

0.61

0.39

473

17.2

163

21.7

464

45.7

61.8

1.53

58.8

3.90

9.73

0.55

36.6

1.51

0.45

0.05

2.82

0.65

9.95

6.06

165

11.6

Ran

ge0.

830.

920.

381.

0645

849

214

718

841

249

860

.163

.054

.862

.69.

2010

.334

.937

.70.

420.

502.

433.

574.

9816

.715

217

3LR

60.

880.

200.

390.

0524

595

.310

412

.021

028

.472

.02.

5493

.76.

2914

.31.

0219

.20.

960.

570.

232.

130.

456.

011.

6614

430

.9R

ange

0.59

1.15

0.34

0.46

161

394

88.8

121

161

233

69.6

75.8

86.3

100

13.2

15.5

17.7

20.2

0.29

0.81

1.51

2.61

3.62

8.29

114

185

LR7

0.16

0.08

0.11

0.02

33.5

10.3

34.5

7.69

86.6

18.7

65.2

6.49

81.9

10.4

14.6

1.62

11.2

1.04

0.06

0.02

1.06

0.13

1.77

0.38

50.1

9.18

Ran

ge0.

030.

260.

070.

1415

.946

.519

.041

.750

.011

151

.472

.061

.495

.811

.416

.89.

7012

.50.

030.

080.

841.

271.

082.

1334

.061

.9LR

85.

421.

671.

430.

3126

843

.227

232

.050

494

.910

010

.612

211

.715

.01.

0625

.02.

962.

890.

785.

241.

889.

200.

6663

.410

.9R

ange

3.35

8.06

1.02

1.82

226

345

225

313

400

663

92.7

119

110

143

13.9

16.4

21.6

29.1

1.83

3.95

3.13

8.22

8.02

9.91

46.1

74.6

LR9

0.31

0.07

0.70

0.19

59.1

12.2

64.8

6.08

109

13.5

59.9

5.21

65.7

1.36

9.30

0.29

13.2

1.14

0.31

0.11

2.89

0.12

2.91

0.74

67.1

12.1

Ran

ge0.

210.

360.

520.

9045

.774

.656

.068

.990

.712

156

.267

.564

.467

.69.

009.

7011

.914

.30.

200.

422.

732.

992.

343.

9953

.682

.5LR

100.

180.

060.

160.

0229

.94.

4834

.53.

0273

.14.

4852

.47.

2672

.012

.111

.81.

9810

.60.

780.

110.

041.

180.

222.

330.

4846

.94.

40R

ange

0.10

0.23

0.14

0.17

23.6

33.8

30.6

37.3

69.8

79.4

42.7

58.0

58.0

82.5

9.60

13.5

9.50

11.2

0.07

0.15

1.06

1.50

1.66

2.77

41.3

51.9

Har

bor

2.49

1.97

1.21

0.85

254

142

161

80.1

409

237

72.5

14.1

86.7

18.7

12.7

2.05

22.8

6.87

1.22

0.95

3.88

2.07

7.36

3.29

91.4

37.1

Ran

ge0.

108.

060.

144.

6623

.663

730

.642

969

.812

6042

.711

954

.814

39.

0016

.49.

5037

.70.

073.

951.

069.

681.

6617

.241

.318

5

N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167 161

ately contaminated sediments. Only the ‘‘background value” layer,the deepest one in profile LR7, consists of essentially uncontami-nated sediments.

The Igeo values of Cd and Pb are similar in their distribution andwill be discussed together. In the most harbor profiles, the valuesare between 2 and 3, corresponding to moderately to heavily con-taminated sediments (Table 3). For both Cd and Pb, Igeo values arevery low (0–1) in all sediment layers of profile LR10, representinguncontaminated to moderately contaminated sediments. Very lowvalues for Pb are also found in some sediment layers of profile LR9(both profiles are located in front of the container terminal ofRijeka port, east of the Rjecina River Delta). Similar low valuesare observed in profile LR7, representing marine sediments in thisregion.

The Igeo values of Cu and Zn are also similar in behavior andwill also be discussed together. In most harbor profiles, the valuesare between 1 and 2, indicating moderately contaminated sedi-ments (Table 3). It is interesting to note a trend of improving Cuand Zn pollution levels in the first two shallowest sediment layers.At some locations, the present situation is similar to that of thedeepest sediment layers, representing the earliest portion of theperiod examined, while the intervening time was a period of in-creased pollution for these two elements. In some sediment layersfrom this period, the Igeo values for Cu and Zn are between 2 and 3,corresponding to moderately to heavily contaminated sediments.In profiles LR9 and LR10 (near the container terminal), the Igeo val-ues of Cu and Zn are very low (between 0 and 1), indicating uncon-taminated to moderately contaminated sediments, as found inprofile LR7 (from the clean marine environment). Some sedimentlayers even have negative Igeo values, indicating a practicallyuncontaminated environment.

Cr and Ni have negative (<0) Igeo values in almost all sedimentlayers of all sampling profiles in the harbor and open sea (Table 3),indicating a practically uncontaminated environment, except inprofile LR8. In the sediment layers of profile LR8, Igeo values forCr and Ni vary between 0 and 1, indicating uncontaminated tomoderately contaminated sediments. This slight contamination isdue to the municipal-sewage outflow near profile LR8 that wasabandoned in 1994.

Co has negative (<0) Igeo values in all sediment layers of allsampling profiles (Table 3), indicating that the sediments of Rijekaharbor and its surroundings are practically uncontaminated withCo. Igeo values for As are between 0 and 1, indicating uncontami-nated to moderately contaminated sediments in most sedimentlayers of all harbor sampling profiles. In several sediment layersof these profiles and in all sediment layers of profiles LR9 andLR10 (located in front of the container terminal) and LR7 (the mar-ine profile), the values are negative (<0), indicating practicallyuncontaminated sediments.

Ag has the highest Igeo values after Hg, which means that it isthe second-heaviest pollutant in Rijeka harbor among the studiedelements (Table 3). The Igeo values for Ag are highest (4–5) in pro-file LR8 (near the abandoned sewage outflow), indicating a heavilyto extremely contaminated environment. In most sediment layers(except the deepest one) of profiles LR1–LR3 from the inner-harborbasin, Igeo values for Ag are between 3 and 4, indicating heavilycontaminated sediments. A slightly better situation, with valuesbetween 2 and 3 (moderately to heavily contaminated) and be-tween 1 and 2 (moderately contaminated), is found in the threeprofiles (LR4–LR6) located in front of the oil refinery. In the sedi-ment layers of profile LR9, which is located close to the Rjecina Riv-er Delta in front of the container terminal, the Igeo values for Agare between 1 and 2 (moderately contaminated), while in profileLR10 (several hundred meters eastward), they are between 0 and1 (uncontaminated to moderately contaminated). The sedimentslayers of the marine-sediment profile (LR7) have negative or

Page 9

Tabl

e5

Conc

entr

atio

nsof

met

als

(mg/

kg)

inse

dim

ents

from

the

wor

ldan

dA

dria

tic

harb

ors.

Gra

insi

zeH

gC

dPb

Cu

ZnN

iC

rA

gM

nA

s

Fou

rh

arbo

rsin

Gu

am<1

mm

0.00

3–0.

740.

06–2

.18

0.38

–123

2.2–

181

3.3–

552

2.1–

102

3.3–

52.7

<0.1

6–1.

80.

98–1

7.0

Den

ton

etal

.(20

05)

Two

har

bors

,Mex

ico

Bu

lk0.

85–6

.18

23.2

–101

9.5–

14,5

5245

.8–1

962

18.3

–135

253–

1093

Hu

erta

-Dia

zet

al.(

2008

)K

aoh

siu

ng

Har

bor,

Taiw

an<1

mm

0.1–

8.5

0.1–

6.8

9.5–

470

5–94

652

–136

90.

2–90

0C

hen

etal

.(20

07)

Har

bor

ofC

euta

.Spa

inB

ulk

10–5

165–

865

29–6

958–

671

13–3

814–

4261

–332

Gu

erra

-Gar

cia

and

Gar

cia-

Gom

ez(2

005)

Nap

les

har

bor

(Ita

ly)

<2m

m0.

01–1

390.

01–3

19–3

083

12–5

743

17–7

234

4–36

27–

1798

1–11

21Sp

rovi

eri

etal

.(20

07)

Port

ofB

agn

oli

Bu

lk0.

01–9

.30.

01–3

.24

52–8

960.

5–12

691

–231

30.

01–5

2.5

4–54

457–

5947

0.5–

4.0

Rom

ano

etal

.(20

04)

Port

oM

argh

era

(Ven

ice)

<1m

m0.

5–4.

80.

4–4.

050

–131

10–9

297

–524

8.0–

5914

–66

0.02

–0.8

533

8–71

3Zo

nta

etal

.(20

07)

Gu

lfof

Trie

ste

Bu

lk0.

06–2

3.3

10–5

033

–145

26–1

5026

–146

336–

979

Cov

elli

etal

.(20

06)

Nor

then

Adr

iati

cse

aB

ulk

0.01

–0.2

38.

2–39

26–1

1242

–134

200–

800

Fabb

riet

al.(

2001

)Ši

ben

ikh

arbo

rB

ulk

2.18

–15.

00.

39–0

.95

61.7

–469

26.2

–374

83.8

–687

17.1

–55.

220

.1–7

1.2

0.19

–0.4

532

3–87

118

.3–4

1.4

Cu

krov

etal

.(20

08)

Rij

eka

har

bor

Bu

lk0.

10–8

.06

0.14

–4.6

623

.6–6

3730

.6–4

2969

.8–1

260

54.8

–143

42.7

–119

0.07

–3.9

59.

50–3

7.7

Ou

rre

sear

ch

162 N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167

slightly positive values, indicating sediments that are practicallyuncontaminated or slightly contaminated sediments with Ag.

Igeo values for Mo are mostly within categories 0–1 and 1–2(uncontaminated to moderately contaminated sediments) andare rather equally distributed across all locations (Table 3), exceptin most sediment layers of profile LR10 and all sediment layers ofprofile LR7, where negative (<0) values are present, indicatingpractically uncontaminated sediments.

Igeo values for Sn are between 1 and 2 in the majority of pro-files, indicating moderately contaminated sediments (Table 3). Inprofiles LR9 and LR10, located in front of the container terminal,Igeo values for Sn are lower (between 0 and 1), indicating uncon-taminated to moderately contaminated sediments. A similar situa-tion is found in the marine location (LR7) in front of the port,where some sediment layers even have negative (<0) values, indi-cating practically uncontaminated sediments.

Ba generally has low Igeo values at all locations (between 0 and1), indicating uncontaminated to moderately contaminated sedi-ments (Table 2). There are even some sediment layers with nega-tive (<0) values, indicating practically uncontaminatedsediments. The exceptions are several sediment layers of profilesLR4 and LR5 (near the oil refinery), where the values are greaterthan 1, indicating moderately contaminated sediments. This con-tamination probably originates from the oil refinery.

4.3.3. Sediment qualityTable 4 shows the concentrations of selected metals in the sed-

iment of Rijeka harbor, while Table 5 shows the concentrations ofthese metals in sediments from other harbors in the Adriatic andworldwide.

Among all studied elements, Hg is the heaviest pollutant in Ri-jeka harbor. At some locations, Hg concentrations exceed 4 mg/kg,and the average value for the whole harbor is about 2.5 mg/kg.These concentrations are comparable to values present in the portof Venice (Zonta et al., 2007) but much lower than the highest val-ues measured in the port of Naples (Sprovieri et al., 2007) and inthe Gulf of Trieste (Covelli et al., 2006). According to sediment-quality guidelines for metals, the concentrations measured in Rije-ka port correspond to values which have extreme effects accordingto most available criteria (Burton, 2002).

Cd concentrations in Rijeka harbor are comparable to those ofall harbors that we used for comparison (Tables 4 and 5). However,maximum concentrations of Cd are higher than in other Adriaticport sediments. Regarding toxicity, Cd concentrations in Rijekaharbor are close to the threshold level according to most guide-lines. At some locations, Cd concentrations are around the mid-range-effect level (Burton, 2002). In the E.U. Water FrameworkDirective (2000/60/EC), cadmium, mercury and their compoundsare classified as priority hazardous substances and are requiredto be controlled for progressive reduction of discharge, emissionsand losses (Chon et al., 2010).

Pb values in Rijeka harbor are comparable to those in most har-bors that we used for comparison (Tables 4 and 5). According tomost available sediment-quality criteria, average Pb values in Rije-ka harbor as a whole (about 250 mg/kg) correspond to midrange orextreme effects, depending on the criteria. The highest Pb valuesfound in the harbor are well above those corresponding to extremeeffects (Burton, 2002). In the E.U. Water Framework Directive(2000/60/EC), lead, nickel and their compounds are classified aspriority substances and are required to be controlled for progres-sive reduction of discharge, emissions and losses (Chon et al.,2010).

Cu concentrations in Rijeka harbor are lower than in most ofthe compared ports (Tables 4 and 5). According to mostsediment-quality criteria, the average value for the whole portarea (about 160 mg/kg) corresponds to midrange or extreme

Page 10

Table 6Basic statistical parameters for all analyzed elements.

Mean Geometric Median Minimum Maximum Std. dev. Skewness Kurtosis

Hg 2.20 1.18 1.42 0.03 8.06 1.99 0.89 0.10Li 38.9 38.3 38.1 22.5 55.1 6.5 �0.04 0.38Be 0.92 0.91 0.90 0.50 1.20 0.16 �0.05 �0.12B 40.6 39.8 40.5 22.0 64.0 8.13 0.21 0.62Na 1.03 0.99 1.00 0.57 1.93 0.27 0.81 1.33Mg 1.40 1.36 1.31 1.10 2.86 0.37 2.94 8.34Al 1.77 1.75 1.75 1.04 2.45 0.27 0.19 0.46P 0.09 0.08 0.07 0.03 0.36 0.06 2.14 5.69S 0.74 0.63 0.77 0.19 1.67 0.38 0.36 �0.68K 0.40 0.39 0.38 0.21 0.70 0.10 1.24 1.75Ca 8.86 8.52 7.66 5.57 17.6 2.70 1.49 1.93V 72.4 70.8 68.0 40.0 139 16.5 1.60 4.46Cr 71.6 70.4 69.6 42.7 119 13.6 1.30 2.98Ti 0.01 0.01 0.01 0.01 0.03 0.01 1.16 0.39Mn 343 338 331 228 478 59.3 0.24 �0.78Fe 2.74 2.72 2.77 1.72 3.36 0.33 �0.90 1.16Co 12.9 12.8 13.3 9.00 16.8 2.09 �0.22 �0.87Ni 86.1 84.4 83.3 54.8 143 17.9 0.90 1.37Cu 145 116 148 19.0 429 85.9 0.70 0.87Zn 369 281 383 50.0 1260 246 0.97 1.85Ga 6.04 5.98 5.98 3.79 8.01 0.84 �0.02 0.04As 21.4 20.0 22.0 9.50 37.7 7.50 0.13 �0.65Se 1.05 0.96 1.00 0.30 2.50 0.43 0.92 1.74Rb 31.3 30.7 30.1 17.2 48.0 6.40 0.74 1.01Sr 213 193 163 107 579 111 1.75 2.86Y 11.3 11.3 11.4 7.38 14.5 1.25 �0.42 1.21Zr 0.96 0.91 0.90 0.40 2.30 0.36 1.95 5.50Sc 5.41 5.36 5.35 3.50 7.20 0.76 �0.03 �0.17Pr 2.91 2.89 2.85 2.30 3.80 0.34 0.50 �0.42Gd 2.59 2.58 2.60 1.90 3.20 0.26 �0.02 0.21Dy 2.00 1.99 2.02 1.49 2.49 0.20 �0.07 0.76Ho 0.38 0.38 0.40 0.30 0.50 0.05 �0.82 0.22Er 0.97 0.96 1.00 0.70 1.20 0.10 �0.09 0.13Tm 0.11 0.11 0.10 0.10 0.20 0.03 3.42 10.1Nb 0.37 0.35 0.35 0.20 0.80 0.12 1.23 2.68Mo 3.53 2.93 3.16 0.84 9.68 2.16 1.06 0.72Ag 1.07 0.60 0.82 0.03 3.95 0.97 1.12 0.82Cd 1.07 0.70 0.97 0.07 4.66 0.87 1.47 3.92In 0.07 0.06 0.06 0.02 0.30 0.05 2.23 6.66Sn 6.66 5.56 6.76 1.08 17.2 3.60 0.61 0.85Sb 1.62 1.27 1.59 0.19 4.68 0.99 0.70 0.69Te 0.09 0.08 0.09 0.04 0.16 0.02 0.66 1.21Cs 1.77 1.74 1.71 1.08 2.79 0.36 1.29 1.56Ba 86.3 79.3 73.8 34.0 185 37.4 1.06 0.36La 10.7 10.6 10.5 8.10 14.1 1.44 0.59 �0.38Ce 23.1 22.9 22.8 18.5 30.8 2.91 0.67 �0.10Nd 11.4 11.3 11.2 8.39 14.5 1.28 0.46 �0.03Sm 2.49 2.47 2.45 1.90 3.30 0.28 0.58 0.57Eu 0.60 0.59 0.60 0.50 0.80 0.06 0.53 1.88Tb 0.38 0.37 0.40 0.30 0.50 0.05 �0.73 �0.06Yb 0.71 0.71 0.70 0.50 0.90 0.09 0.04 0.11Re 0.01 0.01 0.01 0.00 0.02 0.00 0.68 0.42Au 73.5 46.8 40.5 5.00 550 84.9 3.51 17.5Tl 0.39 0.37 0.39 0.17 0.69 0.13 0.48 �0.36Pb 227 161 228 15.9 637 152 0.52 �0.01Bi 0.46 0.25 0.45 0.02 1.35 0.36 0.45 �0.45Th 3.27 3.23 3.30 2.50 4.50 0.48 0.38 �0.12U 2.61 2.20 2.10 0.80 9.90 1.76 2.09 5.57

N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167 163

effects, but the locations with the highest Cu concentrations arewell above values corresponding to extreme effects (Burton,2002).

Zn concentrations in Rijeka harbor are comparable to those inmost ports we used for comparison (Tables 4 and 5). The rangeof Zn concentrations in Rijeka harbor is closest to that obtainedin Kaohsiung harbor, Taiwan. The average concentrations in Rijekaport (about 400 mg/kg) are within midrange-effect values, but themost polluted locations exhibit extreme effects (Burton, 2002).

The range of Cr concentrations is closest to those found in theGulf of Trieste and the Northern Adriatic Sea but is somewhat high-er than in the ports of Venice and Guam (Tables 4 and 5). Other

ports we used for comparison have higher Cr concentrations. Aver-age Cr concentrations in Rijeka harbor (about 70 mg/kg) are closeto threshold values according to most sediment-quality criteria.The highest Cr concentrations found in this study exceed thethreshold level but fall below midrange-effect values (Burton,2002).

The range of Ni concentrations is comparable with those fromthe Gulf of Trieste and the Northern Adriatic Sea (Tables 4 and5). Average Ni concentrations in Rijeka harbor (about 86 mg/kg)correspond to midrange or extreme effects, depending on the sed-iment-quality criteria. The locations with the highest Ni concentra-tions exhibit extreme-effect values (Burton, 2002).

Page 11

Table 7Anomalies (extremes and outliers) determined by boxplot method.

Sample Extreme Outlier

LR1-1 – –LR1-2 – –LR1-3 – CuLR1-4 – NaLR1-5 – –LR1-6 – –LR1-7 – –LR2-1 Au –LR2-2 – ReLR2-3 – MoLR2-4 – –LR2-5 – –LR2-6 – –LR3-1 – –LR3-2 – –LR3-3 – –LR3-4 – ULR3-5 – –LR3-6 – –LR3-7 – –LR4-1 – –LR4-2 – –LR4-3 – –LR4-4 Zr, Nb V, Er, InLR4-5 V, Zr, In, U B, Zn, Se, Sr, Nb, Mo, CdLR4-6 – V, Se, Zr, Mo, TeLR5-1 – –LR5-2 – V, SrLR5-3 – NbLR6-1 – –LR6-2 V Mg, Er, Ba, Sm, ULR6-3 – –LR6-4 – –LR6-5 – –LR7-1 – K, Rb, CsLR7-2 Rb K, CsLR7-3 – K, Rb, CsLR7-4 – K, Rb, Cs, SmLR7-5 – K, Rb, CsLR7-6 K, Rb Li, Cs, Ce, Sm, TbLR7-7 – Ca, SrLR8-1 – P, CrLR8-2 P, Cr Ni, Y, AgLR8-3 Cr P, NiLR8-4 – P, CrLR8-5 – Cr, SbLR8-6 – Cr, MoLR8-7 – Cr, NiLR9-1 Mg –LR9-2 Mg CaLR9-3 Mg TeLR9-4 Mg CaLR10-1 – –LR10-2 – –LR10-3 – –LR10-4 – –

Table 8Mean values for all analyzed elements for each cluster.

Element Cluster 1 mean Cluster 2 mean Cluster 3 mean

Hg 3.12 5.68 0.64Li 38.2 37.5 39.7Be 0.87 0.84 0.98B 39.2 38.9 41.9Na 1104 1056 9731Mg 1224 1253 1538Al 1681 1698 18,377P 117 154 57.6S 966 1213 470K 369 389 417Ca 7497 6960 10,197V 69.8 67.4 75.4Cr 73.7 86.9 65.8Ti 13.0 16.0 14.3Mn 318 386 344Fe 2749 2961 2671Co 12.7 14.7 12.5Ni 87.6 109 78.3Cu 178 252 94Zn 461 489 280Ga 5.80 5.84 6.23As 23.9 24.5 19.0Se 1.14 1.22 0.95Rb 29.6 29.4 32.8Sr 172 150 256Y 11.2 11.3 11.4Zr 0.87 0.77 1.07Sc 5.18 5.51 5.52Pr 2.75 2.64 3.07Gd 2.51 2.51 2.66Dy 1.96 2.00 2.03Ho 0.37 0.39 0.38Er 0.96 0.98 0.97Tm 0.10 0.10 0.11Nb 0.38 0.32 0.37Mo 4.59 4.84 2.54Ag 1.75 2.03 0.40Cd 1.53 1.30 0.74In 0.08 0.06 0.08Sn 7.95 8.71 5.32Sb 2.06 2.38 1.14Te 0.08 0.09 0.09Cs 1.64 1.60 1.90Ba 79.4 67.7 95.7La 9.93 9.37 11.46Ce 21.7 20.4 24.6Nd 10.9 10.6 12.0Sm 2.38 2.39 2.58Eu 0.58 0.59 0.61Tb 0.37 0.39 0.38Yb 0.70 0.73 0.71Re 0.01 0.01 0.01Au 0.12 0.15 0.03Tl 0.49 0.47 0.32Pb 257 258 200Bi 0.66 0.76 0.26Th 3.08 3.33 3.36U 2.84 2.39 2.55

164 N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167

Concentrations of As are lower than those in Naples harbor (Ta-ble 4 and 5) but are higher than in the ports of Guam and Bagnoli.Average As concentrations in Rijeka harbor (about 23 mg/kg) aresomewhat higher than threshold-effect values but are lower thanmidrange-effect values according to most sediment-quality crite-ria. According to some sediment-quality criteria, the highest Asvalues in Rijeka harbor correspond to midrange effects (Burton,2002).

4.4. Statistical interpretation of results

4.4.1. Basic statistical parametersBasic statistical parameters for all analyzed elements in bulk

sediments are presented in Table 6. These parameters were

obtained for all 58 analyzed elements in the whole dataset, includ-ing samples from all sediment layers and localities.

4.4.2. Determination of anomalies using the boxplot methodAnomalies (extremes and outliers) were determined by the

boxplot method, and the results are presented in Table 7.In bulk sediments, the most important anomalies are as follows:

a strong Au extreme is present in the shallowest layer of profileLR2; some anomalies of V, Zn, Rb, Sr, Zr, Er, Nb, Mo, Cd, In, Teand U are present in deeper sediment layers of profile LR4, nearthe oil refinery; K, Rb and Cs outliers and extremes are present inall sediment layers of profile LR7, which is the profile with the

Page 12

Table 10Factor scores for all three obtained factors in all studied sediment samples.

Factor 1 Factor 2 Factor 3

LR1-1 �0.28 �0.56 0.55LR1-2 �0.15 �0.04 0.55LR1-3 0.37 �0.60 1.20LR1-4 0.09 �0.01 0.83LR1-5 �0.18 �1.14 0.48LR1-6 �0.16 �0.11 0.47LR1-7 �0.28 �0.53 0.31LR2-1 0.37 �0.40 0.22LR2-2 0.73 �1.17 �0.15LR2-3 0.83 �0.94 0.04LR2-4 0.57 �0.52 �0.15LR2-5 0.22 �0.37 0.00LR2-6 �0.05 �0.65 �0.23LR3-1 0.09 �0.15 0.17LR3-2 0.18 �0.41 0.11LR3-3 0.33 �0.19 0.05LR3-4 1.43 0.34 �0.12LR3-5 0.82 �0.61 �0.51LR3-6 0.00 0.57 0.13LR3-7 �0.86 0.55 �0.01LR4-1 0.40 0.19 �0.52LR4-2 0.58 0.67 �0.39LR4-3 0.91 1.03 �0.53LR4-4 2.56 1.37 �0.55LR4-5 3.47 0.51 �0.91LR4-6 1.62 0.83 �0.44LR5-1 1.25 �0.32 �1.39LR5-2 1.19 �0.98 �1.48

N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167 165

strongest marine influence; P and Cr extremes and outliers arepresent in almost all sediment layers of profile LR8, and Ni andSb outliers are found in some sediment layers of the same profile,probably due to municipal-sewage outflow; and strong Mg ex-tremes are present in all sediment layers of profile LR9, where Caoutliers are also found in some sediment layers.

4.4.3. Q-modality cluster analysisQ-modality cluster analysis was performed on bulk sediments

using data for all 58 analyzed elements. The following three clus-ters were obtained: Cluster 1, including 17 samples; Cluster 2,including 9 samples; and Cluster 3, including 30 samples. Table 8shows the mean values for all analyzed elements for each clusterto illustrate the differences between clusters in concentrations ofparticular elements.

Cluster 1 includes almost all sediment layers from samplingpoints 2 and 3 and the two upper sediment layers of samplingpoints 1 and 8. Locations 1, 2 and 3 are within the inner part ofthe Rijeka Port basin, while location 8 is located at the old RjecinaRiver channel. Moreover, the municipal-sewage outflow aban-doned in 1994 was located near profile 8. Samples from Cluster 1could be characterized as ‘‘slightly polluted”; the concentrationsof most toxic elements are higher than those found in Cluster 3but lower than those found in Cluster 2, which represents the mostpolluted samples. Cluster 2 contains deeper sediment layers thanlocations 1 and 8 and represents the most polluted of the threeclusters. We attribute this result to the municipal-sewage outflownear location 8, which was active for a long time during the past.Only the two shallowest sediment layers from these sites are foundin Cluster 1, indicating that the situation has improved substan-tially since the closing of this pollutant source. Cluster 3 containssamples from all other locations and represents the cleanest envi-ronment of Rijeka Port and its surroundings. Location 7, whosedeepest layer was chosen to represent clean sediment from thisarea, is found in this cluster. Cluster 3 also shows the strongestmarine influence.

4.4.4. Factor analysisFactor analysis was performed on bulk sediments using data for

20 analyzed elements. The most important elements were selectedas representatives for both natural and anthropogenic sources be-cause the goal of factor analysis is to reduce the total amount of

Table 9Factor loadings data. Factors with significant correlations are presented in bold.

Factor 1 Factor 2 Factor 3

Hg 0.26 �0.27 0.81Li �0.15 0.90 0.08Be 0.38 0.81 �0.22Al �0.13 0.95 0.06P 0.26 �0.15 0.84Cr 0.29 0.25 0.88Ni �0.03 0.29 0.91Cu 0.59 �0.17 0.71Zn 0.92 �0.05 0.27As 0.89 �0.10 0.19Sc �0.09 0.92 0.23Ag 0.29 �0.21 0.86Cd 0.84 �0.08 0.22Sn 0.78 �0.05 0.34Sb 0.82 �0.14 0.37La �0.05 0.86 �0.31Ce �0.08 0.90 �0.35Eu 0.03 0.82 0.07Pb 0.94 0.04 0.09U 0.86 0.15 �0.05Expl.Var 6.12 5.82 4.96Prp.Totl 0.31 0.29 0.25

data and present the data in the simplest way. Three factors wereextracted, and the data were varimax normalized. The estimatedmodel explained 84.49% of the total data. Table 9 presents the fac-tor loadings. Factor 1 has excellent correlations with Zn and Pb andvery good correlations with As, Cd, Sn, Sb and U. This factor isanthropogenic, strongly related to Rijeka port activities and partlyto the oil refinery; the highest factor scores are observed in sedi-ment layers of profiles LR2–LR5, which are located mostly withinthe port basin and close to the oil refinery. Table 10 presents factorscores for all three factors in all sediment samples. Notably, thehighest scores for Factor 1 in some of the profiles are present inthe deeper sediment layers, indicating much higher port activityin the past. Factor 2 has very good and excellent correlations withLi, Be, Al, Sc, La, Ce and Eu and can be explained as a lithogenic

LR5-3 1.65 �0.40 �1.07LR6-1 �0.41 0.85 �0.13LR6-2 0.38 1.92 �0.36LR6-3 �0.56 0.40 �0.20LR6-4 �0.56 1.04 �0.01LR6-5 �0.43 0.75 �0.01LR7-1 �1.14 1.27 �0.59LR7-2 �1.19 1.31 �0.51LR7-3 �1.29 1.30 �0.37LR7-4 �1.18 2.01 �0.41LR7-5 �1.20 1.27 �0.57LR7-6 �1.19 2.77 �0.17LR7-7 �1.22 �0.32 �1.12LR8-1 �0.15 �0.17 1.98LR8-2 �0.14 0.16 2.92LR8-3 �0.07 0.35 2.88LR8-4 �0.47 �0.53 2.14LR8-5 0.63 �0.64 1.76LR8-6 0.61 �0.92 1.72LR8-7 �0.04 �0.15 1.71LR9-1 �0.83 �1.49 �0.95LR9-2 �0.79 �1.31 �0.91LR9-3 �0.48 0.65 �0.91LR9-4 �0.86 �1.07 �0.96LR10-1 �1.27 �2.55 �1.28LR10-2 �1.21 �1.95 �1.04LR10-3 �1.27 �0.42 �0.66LR10-4 �1.35 �0.43 �0.62

Page 13

166 N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167

factor, indicating terrigenous influence. Higher values of this ter-rigenous factor are present in some sediment layers of several pro-files, possibly indicating periodic influence of terrigenous materialin the Bay of Rijeka. Factor 3 has an excellent correlation with Niand very good correlations with Hg, P, Cr, Cu and Ag. This factoris a typical anthropogenic factor, strongly indicating pollution frommunicipal-sewage waters. The highest scores for Factor 3 are pres-ent in profile LR8, located near the municipal-sewage outflow. Thehighest value is present in layer 2 of this profile, while the most re-cent layer exhibits a lower value, indicating that municipal-sewagepollution has decreased since this sewage outflow was abandonedin 1994. Somewhat increased values of Factor 3 are also found inprofile LR1, especially in deeper sediments layers; this patterncan also be explained by municipal-sewage pollution.

5. Conclusion

For the first time, depth profiles of Rijeka harbor were examinedto evaluate anthropogenic inputs during recent decades. Below, wesummarize our conclusions based on our analysis and a statisticalevaluation of Rijeka port sediments.

The fine fraction generally predominates (>50%) in sedimentcolumns. Coarser sediments are located mostly in front of the Rije-ka oil refinery and near the Rjecina River Delta.

Estimated annual sedimentation rates are 5–6 mm for the in-ner-harbor area (LR1–LR6), 6–7 mm in the Sušak basin (LR8–LR10), and 3–4 mm at the reference station 2 km from the harbor(LR 7).

Based on enrichment factors (EF), we conclude that Hg is theheaviest pollutant in the Rijeka harbor area, followed by Ag. EF val-ues of almost all toxic elements show a decreasing trend during thelast 25 years, indicating an improving situation. Exceptions includeAg and Ba, which show a constant increase in EF values, and Pb,whose EF values are about the same today as in the past. The spa-tial distribution of EF values shows changing metal inputs in theport environment in recent decades.

Geoaccumulation-index (Igeo) values confirm that Hg is theheaviest pollutant, followed by Ag. According to Igeo values, Hgand Ag concentrations at most locations in Rijeka harbor indicatemoderately to heavily contaminated sediments. Concentrationsof Cu, Zn, Mo and Sn at most locations are characteristic of moder-ately contaminated sediments. For all other toxic metals in Rijekaharbor, the sediment samples are uncontaminated or uncontami-nated to moderately contaminated.

The potential toxicity of some metals was evaluated in compar-ison with available sediment-quality guidelines (Burton, 2002). Weconclude that Hg, Pb, Cu and Ni can cause extreme or midrange toextreme effects. Concentrations of Cd and Zn are near the mid-range-effect level, while Cr and As have concentrations aroundthe threshold-effect level.

The boxplot method indicates the following important anoma-lies: a strong Au extreme is present in the shallowest sedimentlayer of profile LR2; some anomalies of V, Zn, Rb, Sr, Zr, Er, Nb,Mo, Cd, In, Te and U are present in deeper sediment layers of profileLR4, close to the oil refinery; P and Cr extremes and outliers arepresent in almost all sediment layers of profile LR8, and Ni andSb outliers are present in some sediment layers of the same profile,probably due to municipal-sewage outflow.

Using a Q-modality cluster analysis, the sediments of Rijekaharbor can be grouped into three clusters. The first cluster contains17 samples and represents slightly polluted sediments. This clustermostly includes samples from the inner basin of Rijeka harbor andfrom location LR8, which is near the abandoned sewage outflow.The second cluster represents the most polluted sediments andcontains nine samples, mostly from deeper sediment layers of loca-tions LR1 and LR8. The third cluster contains 30 samples and rep-

resents the cleanest environment of the Rijeka port and itssurroundings.

Three factors were extracted by factor analysis. Factor 1 is ananthropogenic factor correlated with Zn, Pb, As, Cd, Sn, Sb and U.The anthropogenic factor is strongly related to port activities in Ri-jeka harbor and partly to the oil refinery. Higher scores for Factor 1in the deeper sediment layers indicate much higher port activity inthe past. Factor 2 is a lithogenic factor, indicating a terrigenousinfluence. This factor shows the periodic influence of terrigenousmaterial in the Bay of Rijeka. Factor 3, which is correlated withNi, Hg, P, Cr, Cu and Ag, is another anthropogenic factor, associatedwith municipal sewage. Its highest values are found at locationsnear the abandoned sewage outflow.

We conclude that the pollution situation in Rijeka harbor iscomparable with the conditions found in most of the Adriaticand in most harbors around the world. There is an obvious improv-ing trend during the last 25 years with decreasing concentrationsof the most toxic elements, except for Ag and Ba. Despite this gen-eral improving trend, we recommend future monitoring of portsediments in Rijeka harbor to track the further development ofits contamination status.

Acknowledgements

We gratefully acknowledge the financial support of the Port ofRijeka Authority and the Ministry of Science, Education and Sportsof the Republic of Croatia under Projects 098-0982934-2720,‘‘Interactions of trace metals in aquatic environments”, and 098-0982934-2713, ‘‘Radionuclides and trace elements in environmen-tal systems”. We thank Ivan Grabar and Momir Milunovic forassistance with sampling.

References

Burton, J.G.A., 2002. Sediment quality criteria in use around the world. Limnology 3,65–75.

Chen, C.W., Kao, C.M., Chen, C.F., Dong, C.D., 2007. Distribution and accumulation ofheavy metals in the sediments of Kaohsiung Harbor, Taiwan. Chemosphere 66,1431–1440.

Chon, H.S., Ohandja, D.G., Voulvoulis, N., 2010. Implementation of EU WaterFramework Directive: source assessment of metallic substances at catchmentlevels. J. Environ. Monitor. 12, 36–47.

Councell, T.B., Duckenfield, K.U., Landa, E.R., Callender, E., 2004. Tire-wear particlesas a source of zinc to the environment. Environ. Sci. Technol. 38, 4206–4214.

Covelli, S., Fontolan, G., Faganeli, J., Ogrinc, N., 2006. Anthropogenic markers in theHolocene stratigraphic sequence of the Gulf of Trieste (northern Adriatic Sea).Mar. Geol. 230, 29–51.

Cukrov, N., Cuculic, V., Kwokal, Z., 2008. Ecotoxic metals in water and sediment ofthe southeastern part of the of Šibenik harbor, Croatia. In: Jolic, N. (Ed.), ThirdInternational Conference on Ports and Waterways – POWA 2008. Fakultetprometnih znanosti, Dubrovnik, Croatia, pp. 278–286.

Cukrov, N., Mlakar, M., Cuculic, V., Barišic, D., 2009. Origin and transport of 238U and226Ra in riverine, estuarine and marine sediments of the Krka River, Croatia. J.Environ. Radioactiv. 100, 497–504.

Davis, J.C., 1986. Statistics and Data Analysis in Geology, second ed. John Wiley andSons, New York, Chichester, Brisbane, Toronto, Singapore.

Denton, G.R.W., Concepcion, L.P., Wood, H.R., Morrison, R.J., 2005. Trace metals insediments of four harbours in Guam. Mar. Pollut. Bull. 50, 1133–1141.

Dubravic, E., 2001. Rijecka Luka. Muzej Grada Rijeke, Rijeka.Fabbri, D., Gabbianelli, G., Locatelli, C., Lubrano, D., Trombini, C., Vassura, I., 2001.

Distribution of mercury and other heavy metals in core sediments of thenorthern Adriatic Sea. Water Air Soil Pollut. 129, 143–153.

Fichet, D., Radenac, G., Miramand, P., 1998. Experimental studies of impacts ofharbour sediments resuspension to marine invertebrates larvae: bioavailabilityof Cd, Cu, Pb and Zn and toxicity. Mar. Pollut. Bull. 36, 509–518.

Folk, R.L., 1954. The distinction between grain size and mineral composition insedimentary rock nomenclature. J. Geol. 62, 344–356.

Guerra-Garcia, J.M., Garcia-Gomez, J.C., 2005. Assessing pollution levels insediments of a harbour with two opposing entrances. Environmentalimplications. J. Environ. Manage. 77, 1–11.

Halamic, J., Peh, Z., Bukovec, D., Miko, S., Galovic, L., 2001. A factor model of therelationship between stream sediment geochemistry and adjacent drainagebasin lithology, Medvednica Mt., Croatia. Geologia Croatica 54, 37–51.

Huerta-Diaz, M.A., Delgadillo-Hinjosa, F., Hernandez-Ayon, M., Segovia-Zavala, J.A.,Garcia-Esquivel, Z., Lopez-Zarate, H., Siqueiros-Valencia, A., Galindo-Bect, S.,2008. Diagnosis of trace metal contamination in sediments: the example of

Page 14

N. Cukrov et al. / Marine Pollution Bulletin 62 (2011) 154–167 167

Ensenada and El Sauzal, two harbors in Baja California, Mexico. Mar. Environ.Res. 66, 345–358.

Kaufman, L., Rousseeuw, P.J., 1990. Finding Groups in Data: An Introduction toCluster Analysis. Wiley, New York.

Long, E.R., Robertson, A., Wolfe, D.A., Hameedi, J., Sloane, G.M., 1996. Estimates ofthe spatial extent of sediment toxicity in major US estuaries. Environ. Sci.Technol. 30, 3585–3592.

Lough, G.C., Schauer, J.J., Park, J.S., Shafer, M.M., Deminter, J.T., Weinstein, J.P., 2005.Emissions of metals associated with motor vehicle roadways. Environ. Sci.Technol. 39, 826–836.

Müller, G., 1979. Schwermetalle in den Sedimenten des Rheins-Veranderungen Seit1971. Umschau 79, 778–783.

Pravdic, V., 1995. The chemical-industry in the Croatian Adriatic region –identification of environmental-problems, assessment of pollution risks, andthe new policies of sustainability. Sci. Total Environ. 171, 265–274.

Reimann, C., Filzmoser, P., Garrett, R.G., 2005. Background and threshold:critical comparison of methods of determination. Sci. Total Environ. 346, 1–16.

Romano, E., Ausili, A., Zharova, N., Magno, M.C., Pavoni, B., Gabellini, M., 2004.Marine sediment contamination of an industrial site at Port of Bagnoli, Gulf ofNaples, Southern Italy. Mar. Pollut. Bull. 49, 487–495.

Sanders, C.J., Santos, I.R., Silva, E.V., Patchineelam, S.R., 2006. Mercury flux toestuarine sediments, derived from Pb-210 and Cs-137 geochronologies(Guaratuba Bay, Brazil). Mar. Pollut. Bull. 52, 1085–1089.

Sprovieri, M., Feo, M.L., Prevedello, L., Manta, D.S., Sammartino, S., Tamburrino, S.,Marsella, E., 2007. Heavy metals, polycyclic aromatic hydrocarbons andpolychlorinated biphenyls in surface sediments of the Naples harbour(southern Italy). Chemosphere 67, 998–1009.

StatSoft Inc., 2001. Statistica for Windows [software]. Tulsa, StatSoft.Sutherland, R.A., 2000. Bed sediment-associated trace metals in an urban stream,

Oahu, Hawaii. Environ. Geol. 39, 611–627.Thevenot, D.R., Moilleron, R., Lestel, L., Gromaire, M.C., Rocher, V., Cambier, P.,

Bonte, P., Colin, J.L., de Ponteves, C., Meybeck, M., 2007. Critical budget of metalsources and pathways in the Seine River basin (1994–2003) for Cd, Cr, Cu, Hg,Ni, Pb and Zn. Sci. Total Environ. 375, 180–203.

Tukey, J.W., 1977. Exploratory Data Analysis.Wei, T.Y., Chen, Z.Y., Duan, L.Y., Gu, J., Saito, Y., Zhang, W.G., Wang, Y.H., Kanai, Y.,

2007. Sedimentation rates in relation to sedimentary processes of the YangtzeEstuary, China. Estuar. Coast Shelf Sci. 71, 37–46.

Zonta, R., Botter, M., Cassin, D., Pini, R., Scattolin, M., Zaggia, L., 2007. Sedimentchemical contamination of a shallow water area close to the industrial zone ofPorto Marghera (Venice Lagoon, Italy). Mar. Pollut. Bull. 55, 529–542.