Environmental geochemistry of the polymetallic ore deposits : Case studies from the Rude and the Sv . Jakob historical mining sites , NW Croatia

This paper presents the results of sampling surveys carried out in order to evaluate the environmental impact of the Rude and the Sv. Jakob historical mining sites, NW Croatia. The studied polymetallic ore deposits differ in their mineralogical and geochemical features as well as in the host rock lithology. The Rude Fe-Cu-Pb-Zn-Ba deposit is hosted by Permian siliciclastic sediments. Siderite, haematite, galena, sphalerite, chalcopyrite, pyrite, barite and gypsum are the major ore minerals. The Sv. Jakob Pb-Zn deposit occurs in the Middle Triassic dolostone. The most abundant ore minerals are galena, sphalerite and pyrite. Although the deposits represent potential sources of numerous toxic metals, no pollution of the drainage streams and associated stream sediments was recorded. The studied mining sites are characterized by high carbonate/sulfi de ratios responsible for the alkaline character of the drainage streams (pH=7.3–8.7). Consequently, the mining sites have very low potential for both generation of acid mine drainage and for leaching of heavy metals into the drainage systems. Furthermore, the study revealed that the populated areas (stream waters with decreased redox potential, increased organic matter content, high NO3, NH4 and PO4 concentrations; stream sediments enriched in exchangeable Pb and Zn) and the Sava river alluvium (overfl owing streams enriched in Hg) represent a greater environmental threat than the investigated polymetallic ore deposits.


INTRODUCTION
The polymetallic ore deposits represent the geochemical anomalies of numerous potentially toxic metals such as lead, copper, cadmium, mercury, arsenic and barium.Although the primary minerals of these elements (mostly sulfi des) do not represent a signifi cant environmental threat, their weathering under surface conditions may mobilize heavy metals making them biologically available.The environmental be-haviour of the polymetallic ore deposits varies according to the ore and gangue mineralogy, major and trace element geochemistry and the host rock lithology (e.g.MASCARO et al., 2001;FOLEY, 2002;HUANG et al., 2010;KIRSCH-BAUM et al., 2012).
The presented study focused on the two polymetallic ore deposits, at Rude, in the Samoborska Gora Mts., and Sv.Jakob, (Medvednica Mts.), in the vicinity of Zagreb, NW Croatia.The deposits differ in their mineralogical and geo-chemical features as well as in the host rock lithology.The Rude Fe-Cu-Pb-Zn-Ba deposit is hosted by Permian siliciclastic sediments.Siderite, haematite, galena, sphalerite, chalcopyrite, pyrite, barite and gypsum are the principal ore minerals (PALINKAŠ et al., 2010).The Sv. Jakob Pb-Zn deposit occurs in the Middle Triassic dolostone.The most abundant ore minerals are galena, sphalerite and pyrite (ŠINKOVEC et al., 1988).Both deposits are drained by surface streams which could transport potentially toxic metals downstream.
The Rude deposit is drained by the Rudarska Gradna creek.The geochemical and mineralogical study undertaken by ČOVIĆ (2003) to evaluate heavy metal levels in overbank sediments of the Žumberak and Samoborsko Gorje Mts., revealed an increased content of numerous metals (Hg, Cu, Pb, Ba, As, Cr and Fe) in samples from the Rudarska Gradna catchment.
Soils developed above the Sv.Jakob historic mining site were found to contain increased concentrations of Pb, Zn and Cd, related to the oxidation of primary sulfi des in the near-surface weathering environment (DURN et al., 1998).The deposit is located only 300 m north from the Veliki potok creek headwater, and there is a strong possibility that the creek might transport signifi cant concentrations of heavy metals downstream.
The primary objective of this study was to determine the environmental impact of the historical mining sites on the surrounding ecosystems, including the streams that drain them.Stream water chemistry and stream sediment composition are discussed, as well as the geochemical behaviour of metals during the weathering processes in the two different types of polymetallic ore deposits.

MINERALIZATION
The Rude Fe-Cu-Pb-Zn-Ba deposit consists of two ore types, an epigenetic, hydrothermal vein-type beneath a stratiform, SEDEX type.Epigenetic, epi-mesothermal, quartz-siderite veins with chalcopyrite, galena, sphalerite and barite, crosscut the Middle-Upper Permian clastic rocks.The epigenetic part of the deposit represents a feeder zone of the SEDEX, stratiform ore lenses formed in an evaporitic pond-lagoon by sedimentation of gypsum-anhydrite, haematite, siderite and barite.Laterally, epigenetic, epithermal barite-galena veins intersect Upper Permian coarse-grained sandstones.The deposit is covered by Lower Triassic variegated clastics (PALINKAŠ et al., 2008;ŠINKOVEC, 1971).Mining activities in the area began in the early 13 th century with the exploitation of copper.The production of iron ore from siderite started in 1850 and lasted less than 10 years.According to PALINKAŠ et al. (2010) the Rude deposit is declared as a prototype of the Permian siderite-polysulphide-barite deposits produced during the early stage of intracontinental rifting.
The deposit is drained by the Rudarska Gradna creek (Fig. 1), which originates approximately 1.8 km southwest of the deposit.Generally, it runs southwest-northeastwards for about 16 km.In the upper part of the course, the creek fl ows through a rural area and it is fed by several small streams.After its confl uence with the Lipovečka Gradna creek, it fl ows through the town of Samobor.The Gradna creek joins the Sava River in the vicinity of Gradna village.
The Sv. Jakob Pb-Zn deposit is hosted by non-metamorphosed Middle Triassic dolostone (PALINKAŠ et al., 2008), with the ore occurring in the form of veinlets and lenses.Galena is the principal ore mineral, and it is coarse grained.The vein-type mineralization, beside galena, comprises minor sphalerite, pyrite, quartz and calcite (ŠINKOVEC et al., 1988), and was actively mined during the 17 th century.According to the epigenetic character of the ore mineralization, a simple ore paragenesis without copper minerals, and the fl uid inclusions and sulfur isotope data, the deposit is classifi ed as a Mississippi Valley type of Pb-Zn deposits formed during the stage of advanced intracontinental rifting in the Middle Triassic (PALINKAŠ et al., 2008;ŠINKOVEC et al., 1988;DURN et al., 1999;BOROJEVIĆ ŠOŠTARIĆ, 2004).
The deposit is located at 830 m above sea level, approximately 300 m north of the Veliki Potok creek headwater.The Veliki Potok creek runs north-south for about 12 km, and is fed in its upper part by the Mali Potok creek and several small streams.The central and lower part of the creek fl ows through the urban area of Zagreb.In the lowermost part of the course, about 500 m upstream from its entry point into the Sava River; the Veliki Potok has its confl uence with the Vrapčak creek (Fig. 1).The Veliki Potok originates in the Palaeozoic parametamorphic complex (Fig. 1) composed of greywackes, siltstones, limestones, dolostones, muscovitechlorite and muscovite-quartz schists (ŠIKIĆ et al., 1978;ŠIKIĆ et al., 1979).According to several authors (ĐURĐANOVIĆ, 1973;BELAK et al., 1995) some protoliths from the parametamorphic complex are of Triassic age.Continuing downstream, the creek is entrenched in the Cretaceous sedimentary rocks predominantly represented by fi ne-grained clastics.The orthometamorphic complex, predominantly composed of greenschist and metamorphosed gabbro and diabase, is referred to as being of Devonian-Carboniferous age.Further, the creek fl ows through the Neogene sediments represented mostly by sandstones, marls and limestones.The lowermost part of the creek dissects the Sava river alluvial sediments (ŠIKIĆ et al., 1978;ŠIKIĆ et al., 1979).

SAMPLES AND METHODS
Stream water and sediment samples were simultaneously collected from the Rudarska Gradna creek (n=12), the Veliki potok creek (n=11) and the Sava River (n=1) during the sampling campaign in October 2010.The sampling sites are shown on Figure 1 and listed in Table 1.
Collection and sample handling were conducted in accordance to the standard methodology for the analysis of the selected parameters.Water samples were collected using two half-litre polyethylene (PET) bottles, one for the analysis of anions (SO 4 2-, PO 4 3-, Cl -, NO 3 -) and ammonia, and other, which was pre-acid washed (10%, v/v, nitric acid 65%), for trace metal analysis.The bottles were rinsed three times with stream water prior to fi lling with the water sample.In the laboratory, samples for trace metal analysis were fi ltered through 0.45 mm pore cellulose syringe fi lters, acidifi ed with HNO 3 to pH<2 and stored at 4°C.Another bottle was frozen at -18°C for subsequent analyses of anions and ammonia.
Stream sediment samples were collected and stored in double polyethylene bags.At the laboratory, samples were air dried, disaggregated in an agate mortar, homogenized, and fi nally dry-sieved through stainless-steel screens to the fraction of <125 μm.
Water temperature, pH, redox potential, electrical conductivity and dissolved oxygen (DO) measurements were conducted in situ on unfi ltered water using a portable Hach Lange HQ40d Multimeasure Device.
Nitrate was electrochemically measured using a lowlevel calibrated nitrate-sensitive electrode (Hach Platinum Series combination electrode, model 51920), connected to a volt meter (Hach SensIon 2).To avoid the interference effects of bicarbonate ions, the pH of samples was adjusted to 3 by the addition of an acidic buffer (H 2 SO 4 ).Ammonia content was determined using Hach Intelli-CAL™ ISENH3181 gas-sensing ammonia ion selective el-ectrode (ISE) with a replaceable membrane module, refi llable outer body, double junction reference and built-in temperature sensor.The measuring range is in the interval between 0.007 mg/L and 14,000 mg/L NH 3 -N.The sample preparation procedure required the addition of an alkaline buffer (LiOH) to ensure that the ammonia was essentially de-ionised prior to analysis.
The chloride ion concentration was determined using a selective Radiometer Analytical chloride electrode ISE25Cl.
Sulphate was determined by a titrimetric method.A 0.1 ml volume of methyl orange was added to 100 ml of each water sample.The sample was titrated against 0.1 M HCl to a permanent orange colour.An additional 1 ml of 0.1 M HCl and 25 ml of BaCl 2 solution were then added.Samples were heated to boiling and cooled in tap water.After 30 minutes, 25 ml volume of potassium chromate, 1 drop of aluminum chloride and 1.6 ml 0.1 M KOH solutions were added.After 1 hour, any precipitate was removed by fi ltration and 100 ml of each fi ltrate was titrated against 10 ml 10% potassium iodide solution and sodium thiosulphate using starch indicator.
Total phosphates were analyzed by the ascorbic acid method (APHA Standard Method 4500-P; APHA, AWWA,  , 1998).Absorbance was measured at 880 nm using a Hach DR4000 spectrophotometer.
Acidifi ed sample solutions were used for heavy metal concentration measurements, using the fl ame atomic absorption method for Fe, Mn, Pb, Cu, and Zn and the graphite furnace technique for Hg and Cd measurements (Perkin-Elmer AAnalyst 800).
Exchangeable metals were extracted from the stream sediments (m = 10 g) using 250 ml of 1 M NH 4 OAc solution at pH =7.00.The sediment-solution slurry was shaken for 2 h, and the solution separated from the solid by fi ltration.The addition of NH 4 + in excess to the soil displaces the rapid exchangeable alkali and alkaline cations from the exchange sites of the soil particles.The concentrations of metals in the solution were measured with a Perkin-Elmer AAnalyst 800 Atomic Absorption Spectrometer.
Bulk stream sediment composition was estimated using energy dispersive X-ray fl uorescence (EDXRF).Measurements were made with a W anode and Mo secondary target in orthogonal geometry, with measurement parameters of 40 kV and 35 mA.The irradiation time was 1000 s.X-ray spec-tra were collected with a Si(Li) detector (FWHM=170 eV at 5.9 keV) and were analyzed using QXAS program packagedirect comparison method.IAEA "Lake Sediment" was used as a reference material.
Highly oriented samples of the <2 μm fraction of selected stream sediment samples were prepared for clay mineral identifi cation on air-dried, ethylene-glycol saturated, and heated (at 400 and 550 °C, respectively) samples according to the procedure described by Starkey et al. (1984).Instrumental conditions were 40 kV, 40 mA and constant time 5 s, with step scanning (0.02°2θ).

RESULTS
The chemical characteristics of the water samples from the Rudarska Gradna and Veliki Potok creeks are listed in Table 2.According to the Croatian water quality guidelines (OG 77/98), analyzed samples are classifi ed within Classes I and II, concerning the general physicochemical parameters (pH, conductivity and dissolved oxygen; Fig. 2).According to nitrate and ammonium concentrations, the water samples belong to Classes II and III and Classes I and II, respectively (Fig. 3).Increased concentrations of various forms of nitrogen in rural and urban areas indicate the infl uence of domestic wastewaters, mostly attributed to the discharge of sewage into the creeks or to the drainage from septic tanks.The results for total phosphorus classify all samples within Class I (Fig. 2).Measured concentrations of sulfates and chlorides were considerably lower than the maximum allowable concentrations (MAC) allowed by the Croatian regulations for drinking waters (OG 47/08).
Iron and manganese concentrations in water samples mostly exceed the MAC for drinking waters (Fig. 4).Ele-vated lead content classifi es the analysed samples within Classes IV and V. Lead had its highest water content in urban areas (sampling sites RG-11 and VP-10; Fig. 4) refl ecting an anthropogenic origin.According to their mercury concentrations, the water samples mostly belong to Class IV (Fig. 4).The copper level varies from Class I to Class III.Low concentrations of zinc and cadmium are within the ranges proposed for Class I.
A summary of the NH 4 OAc extracted contents of Fe, Mn, Pb, Zn, Cu, Hg and Cd in stream sediments is given in Table 3 and Figure 4. Iron contents vary from 0.550-2.075mg/kg, Mn content from 13.550-107.775mg/kg, Pb up to 4.275 mg/kg, Zn up to 3.425 mg/kg, Cu from 0.275-8.575mg/kg, and Cd from 6.625-98.250mg/kg.Mercury is mostly below the detection limit.The exceptions are samples VP-3 and VP-10 with Hg concentrations of 0.900 and 1.650 mg/ kg, respectively.
The bulk chemical composition of the stream sediments is summarized in Table 4. Mean K content for the sediments on metamorphic and siliciclastic terrains is comparatively higher than for the sediments on carbonate terrain, refl ecting the lithological contribution of this element.The principal carriers of K are K-feldspars, biotite, micas and clay minerals.As expected, the highest Ca values are recorded for the stream sediments collected on carbonate terrains.The Fe concentrations are mostly lower than the average shale value (Table 4).Sampling site RG-7 near the abandoned wasterock dump of the Rude deposit is the exception.The mean concentrations of Pb, Zn, Cu and As are signifi cantly higher than the average shale concentrations.
According to the XRD patterns, kaolinite and illite represent the most abundant clay minerals in all the analyzed stream sediments.

STATISTICAL TREATMENT
The statistical analyses that applied univariate (basic statistic parameters) and bivariate analysis (correlation matrix) were performed on the analytical results obtained from 24 sampling sites.For each sampling site, 42 descriptors consisting of 5 lithological and 37 geochemical variables were observed and included in the statistical treatment.Lithological variables were defi ned as a percentage of selected lithological units at the sampling site (e.g.ROSE et al., 1970;PEH, 1992;HALAMIĆ et al., 2001).As the bedrock lithologies of the Medvednica and Samoborska Gora Mts.comprise a variety of igneous, sedimentary and metamorphic rocks of different stratigraphic and tectonic settings, a simplifi ed representation of lithology with as little as possible loss of useful information was necessary (HALAMIĆ et al., 2001).For this purpose lithological data were generalized from the geological map of Croatia (scale 1:100.000),Zagreb Sheet (ŠIKIĆ et al., 1978).The study areas comprise only fi ve lithological units which were broadly related to the corresponding stratigraphic nomenclature (Table 1).These units were further utilised as lithological variables which were specifi ed as follows: LIT1 -Triassic carbonate rocks; LIT2 -Permian unmetamorphosed complex, Triassic, Cretaceous and Pliocene clastic sediments; LIT3 -Prolluvium and alluvium; LIT4 -Parametamorphites; and LIT5 -Orthometamorphites.Additionally, the settlement infl uence variable was estimated for all sampling sites.An environmental impact of the settlements is pronounced in both urban and rural areas with no sewage and sewerage facilities.

DISCUSSION
The primary ore mineralization usually is not stable under surface conditions (low temperature and pressure, high activities of water, oxygen and CO 2 ).Weathering products could be summarized within 3 principal groups: 1) Residual primary minerals, resistant to chemical and physical weathering; 2) Secondary minerals, including clay minerals, hydrous Fe-Al-oxides, sulfates and carbonates; and 3) Soluble (mobile) products.The mobility of metals depends mostly on pH and the redox potential of the solution as well as on the possibility for the formation of complexes.Metals may be adsorbed onto Fe-and Mn-hydroxides (HEM, 1977;TESSIER et al., 1985;FENG et al., 2007), clay minerals (RYBICKA et al., 1985), organic matter (DAVIS, 1982;ÖZ-KARAOVA GUNGOR et al., 2010) and amorphous silica (JAMES & MACNAUGHTON, 1977).Co-precipitation with ferric and/or manganese oxyhydroxide can also decrease heavy metal content in the stream waters.
The Rude deposit represents a potential source of lead (galena), zinc (sphalerite), iron (siderite, chalcopyrite, haematite), as well as sulfate (gypsum, anhydrite and oxidized sulfi de species).The ore mineralization of the Sv.Jakob deposit is simple and comprises only galena, sphalerite and pyrite.Numerous geochemical surveys revealed an increased content of heavy metals in the topsoils of urban and pristine areas of the Zagreb region (NAMJESNIK-DEJANOVIĆ, 1994;ROMIĆ & ROMIĆ, 2003;SOLLITTO et al., 2010) as well as in the overbank sediments of the Sava River (ČOVIĆ, 2003;PAVLOVIĆ et al., 2004).As both deposits are drained by the surface streams there is the possibility that they also contribute to the pollution.
According to the principal geochemical features, the Rudarska Gradna creek is divided into 3 segments: 1) the headwater, 2) the segment upstream from the Rude mining site, and 3) the segment downstream from the Rude mining site.
Although the Rudarska Gradna creek rises within the Middle Triassic carbonate rocks, the headwater is characterized by a slightly decreased pH value (RG-1; Fig. 2), refl ecting the strong infl uence of the underlying siliciclastic sediments, mostly Cretaceous shales, sandstones and cherts (Fig. 1).An increased Eh value suggests low organic matter content.The NO 3 -, NH 4 + and PO 4 3-concentrations, as well as the content of all analyzed metals, with exception of lead, are very low (Fig. 4).An increased Pb concentration is attributed to the remobilization of lead from the Middle Triassic sediments under moderately low pH.
The segment between the headwater and the Rude mining site displays a uniform pH value (pH≈8.5)despite the presence of several small tributary streams with low alkalinity (e.g.RG-5; pH=8.12).The progressive decrease in redox potential coincides with the increase in nitrate, ammonia and phosphate concentrations, suggesting a common source of organic matter, nitrogen and phosphorous.The stream water shows a gradual decrease in lead content mainly due to its alkaline character causing the precipitation and/or adsorption of lead to the stream sediments.The sediment analyses revealed an increase in the exchangeable and total lead content downstream (Fig. 4) indicating that both processes, adsorption and precipitation, should be considered.
Water at the abandoned waste-rock dump of the Rude ore deposit is characterized by a slight decrease in pH and an increase in the Eh value.Nitrates, ammonia and phosphates are signifi cantly lower than in the main stream, indicating the diminished infl uence of anthropogenic pollution.Increased concentrations of metals in the water are expected due to the weathering of ore minerals.The weathering of sulfi de-bearing parageneses usually results in a large decrease in pH value.In the case of the Rude deposit, the observed decrease is minor, indicating that weathering of siderite reduces the acidifi cation effect.Namely, according to reactions (1) and (2), weathering of siderite produces a neutral to slightly acid solution as indicated in the reactions below (SKOUSEN et al., 1997): Consequently, the weathering of Fe-minerals under oxidizing conditions results in the precipitation of Fe-hydroxide which represents the predominant phase in the bulk sediment composition (Fig. 4).The decreased content of all exchangeable metals (Fig. 4) suggests pH as a key parameter controlling metal mobility.Under decreased pH, the competition between H + and the dissolved metals for ligands is more signifi cant.It decreases the adsorption abilities of metals, and consequently increases their mobility.Sometimes, with only a drop in a few pH units, the fi xation percentage of metals on sediment particles may range from almost 100% to 0% (GUNDERSEN & STEINNES, 2003;PENG et al., 2009).The segment downstream from the Rude mining site displays a decrease in redox potential and an increase in nitrogen and phosphorous species, suggesting the enhanced infl uence of urbanization on the water quality.The concentrations of iron, manganese and zinc in the water gradually decrease downstream, suggesting the diminishing infl uence of the ore deposit (Fig. 3).The increase in copper concentration is attributed to agricultural activities and the usage of CuSO 4 -based fungicides in the vineyards (HALAMIĆ et al., 2003).The stream sediments, characterized by the progressive increase in the exchangeable and total metal contents downstream (Fe, Mn, Pb, Zn, Cu), indicate that the mobility of metals is controlled by adsorption and precipitation.
According to the geochemical features, the Veliki Potok creek is divided into 2 segments: 1) the upper course fl ows through an uninhabited area and 2) the lower course fl ows through the urban area.The upper segment of the creek is characterized by moderate pH and high redox potential.An increased Eh value suggests low organic matter content.The NO 3 -, NH 4 + and PO 4 3-concentrations, as well as the content of all analyzed metals, with the exception of manganese, are very low (Fig. 3; Fig. 4).An increased Mn concentration might be attributed to the remobilization of manganese from the Middle Triassic sediments.The low heavy metal content in the upper course, as well as a low sulfate concentration, indicates that the Sv.Jakob Pb-Zn-Ag deposit does not represent a potential pollution source for the draining streams.
The lower part of the creek is characterized by an increased pH value especially in the segment which fl ows over alluvial sediment .Increased conductivity of the stream water within the populated area generally points to a higher content of total dissolved solids.The progressive decrease in redox potential coincides with the increase in nitrate, ammonia and phosphate concentrations, suggesting a common source of organic matter, nitrogen and phosphorous, mostly from domestic wastewaters in the urban areas without sewerage facilities.Agricultural activity in the suburbs, including the use of synthetic and organic fertilizers, may also increase the nitrate, ammonia and phosphate contents of the stream waters.The stream sediments display a progressive increase in the exchangeable and total metals contents (Fig. 4).
Statistical analyses applied to the results obtained from both of the study areas allow additional data interpretation.A positive correlation between the measured pH values of the stream water and the presence of Triassic carbonates (variable LIT1; r=0.53) is expected due to the high buffering capacity of carbonate rocks.Whereas the content of several exchangeable metals positively correlates with the pH value (r Pb =0.55, r Zn =0.52, r Cd =0.51), so pH should be considered to be an important factor in the control of metal mobility.
The redox potential (Eh) correlates positively with the variable LIT1 (r=0.59) and negatively with variable LIT3 (r=-0.48).Electrical conductivity of water, the measure for the concentration of total dissolved solids, increases in the siliciclastic terrains (LIT2; r=0.54) and decreases in the orthometamorphites (LIT5; r=-0.58).Conductivity is also increased in the urban/rural areas (r=0.56).Generally, the concentration of metals in water does not show correlation with the lithological variables.The exceptions are mercury and zinc which correlate positively with alluvial/proluviall sediments (LIT3; r Hg =0.82, r Zn =0.66).The increased Hg concentrations in the Sava river alluvium were attributed to the urbanization, industrial and mining activities upstream (NAMJESNIK-DEJANOVIĆ, 1994; PA LIN-KAŠ et al., 1996;HALAMIĆ et al., 2003;PAVLOVIĆ et al., 2004).Exchangeable zinc also correlates with LIT3 (r=0.52).
Whereas both of the investigated ore deposits were situated in the vicinity of the settlements, it was important to determine the infl uence of the urban/rural areas on the quality of stream water and sediments.Increased conductivity of stream waters within populated areas (r=0.56)generally indicates a higher content of total dissolved solids.The populated areas correlate well with the concentrations of nitrate (r=0.91),ammonia (r=0.86) and phosphate (r=0.64)indicating their common source, mostly from domestic wastewaters in the urban and rural areas without sewerage facilities.In the rural area, agricultural activities, including the use of synthetic and organic fertilizers, may also increase the nitrate, ammonia and phosphate contents in the stream waters.Furthermore, the urban/rural areas statistically correlate with the concentration of mercury in stream waters (r=0.60) and with the content of exchangeable lead (r=0.49) and zinc (r=0.82)from the stream sediments.For the appropriate interpretation of the heavy metals origin, it is important to keep in mind that the populated areas are mostly located on the alluvial deposits (LIT3; r=0.43).
As mentioned above, the Hg content in stream waters positively correlates with LIT3 and the urban/rural areas.Furthermore, it displays a negative correlation with the Eh value (r=-0.61)indicating the increased Hg mobility in a reducing, organic rich, environment, probably as a consequence of the Hg complexation with organic ligands.Among the organic ligands, humic substances are the most important compounds naturally occurring in soils, sediments and waters (THURMAN, 1986;GUNGOR, & BEKBOLET, 2010).As humic and fl uvic acids are soluble in dilute alkali solutions, and the pH of all water samples collected over LIT3 is above 8, these types of organic molecules could significantly contribute to the mobilization of mercury from the alluvial sediments.
The concentration of copper in the water samples also correlates negatively with measured Eh values (r=-0.75)pointing to the importance of organic complexes in copper mobilization.According to WU et al., (2002) humate and fl uvate complexes mobilize Cu (II) ions at pH greater than 8. Negative correlation (r=-0.42) between lead in water and the measured redox potential suggests that formation of the Pb-organic complexes might also mobilize lead.

CONCLUSION
Although the Rude and the Sv.Jakob polymetallic ore deposits represent potential sources of numerous heavy metals, pollution of water and stream sediments downstream from the deposits was not recorded.Buffering reactions with the carbonate country rocks (dolostone, limestone), as well as with carbonate ore and gangue minerals (siderite, calcite), are responsible for the alkaline character of the drainage streams.Consequently, the deposits are not considerably weathered and the mobility of metals is signifi cantly reduc ed.
The present study revealed that the populated areas represent a greater environmental threat than the investigated polymetallic ore deposits.Stream waters in the populated areas, especially in the urban and rural areas without sewerage facilities, are characterized by a decreased redox potential (attributed to the high organic matter content), increased conductivity and increased concentrations of nitrate, ammonia and phosphate.Stream sediments in the urban/rural areas have high contents of exchangeable lead and zinc.
The Sava river alluvium, characterized by its high mercury content, represents an additional environmental threat.The accumulation of Hg within the alluvial deposits is attributed to the urbanization, industrial and mining activities upstream.The overlapping of the populated areas (decreased redox potential, high organic matter content) and the Hg enriched alluvium resulted in the increased Hg mobility probably due to the complexation of mercury with the organic ligands.

Figure 1 :
Figure 1: Simplifi ed geological setting of the investigated areas with marked sampling sites along a) the Rude ore deposit and marked sampling sites along the Rudarska Gradna creek (RG-1 -RG-12); b) the Sv.Jakob ore and marked sampling sites along the Veliki Potok creek (VP-1 -VP-12).

Figure 2 :
Figure 2: The physico-chemical parameters of a) the Rudarska Gradna creek, b) the Veliki Potok creek.

Figure 3 :
Figure 3: Variations in nitrate, ammonia and phosphate content along a) the Rudarska Gradna creek, b) the Veliki Potok creek.

Table 1 :
Description and location of the sample sites.

Table 2 :
Stream water analytical data for the Rudarska Gradna and Veliki Potok creeks.

Table 3 :
Exchangeable metals content in the stream sediments from the Rudarska Gradna and Veliki Potok creeks.

Table 4 :
Bulk chemical composition of the stream sediments from the Rudarska Gradna and Veliki Potok creeks.