- Email: [email protected]
Concentrations and leachability of chemical elements in estuarine sulfur-rich sediments, W. Finland
The Science of the Total Environment 284 Ž2002. 109᎐122
Concentrations and leachability of chemical elements in estuarine sulfur-rich sediments, W. Finland ˚ ¨ Pasi PeltolaU , Mats Astrom ˚ Akademi Uni¨ ersity, Abo ˚ 20500, Finland Department of Geology, Abo Received 1 February 2001; accepted 9 April 2001
Abstract Concentrations, distributions and mobility of chemical elements were investigated in reduced sulfur-rich estuarine sediments located in western Finland. The main objective was to determine the possible extent of metal leaching when dredged masses of these sulfur-rich sediments are dumped on the land and thus exposed to air. When dredged, the reduced sulfur in the sediments oxidises resulting in a lowering of pH, which in turn is expected to leach metals. The study area is an artificial lake claimed from the Botnian sea in 1962. In this lake, several mass-kills of fish have occurred, believed partly to be due to dredging. Two sediment samples Ž0᎐50 and 50᎐100 cm. were taken from 39 sampling points in the lake. These samples were leached in aqua regia Ž2:2:2 HNO3rHClrH 2 O. and analysed for Fe, Al, Mg, Ca, K, P, Na, Mn, Zn, Ba, V, Sr, Cr, Ni, Cu, Co, As, Pb, B, Mo and Cd with ICP-AES. Sulfur and organic carbon were analysed with Leco. In a controlled laboratory experiment, the sediments were allowed to oxidise for 1 year while moisturised with deionised water every month. The pH and conductivity were determined in the beginning of the experiment Žreduced state. and in the end Žoxidised state.. In the supernatants in the oxidised states the amount of leached metals ŽNa, Al, Mn, Zn, Sr, Co, Ni, Cu, Cd, Cr, Pb, U, Li, Rb and As. were determined with ICP-MS. The sediments were found to contain low levels of toxic metals but, as expected, high concentrations of sulfur. In the experiment, pH was lowered Ždown to 3.0. and the conductivity increased in all samples due to oxidation and release of metal ions. The extent of leaching varied between 0.03% for As and 12.3% for Na. Critical pH values, at which high amounts of metals begin to leach, were obtained graphically. These values varied between 4.8 ŽNi. and 3.3 ŽCr.. Not all elements were controlled by pH, e.g. Mn correlated well with its aqua regia leachable concentration. In a planned dredging operation in the area some 23 300 t Ž10 500 m3 . Ždry wt.. of sediments will be dredged. The amounts of metals likely to be leached, according to the results from this study, are as follows Žkg.: Al Ž1710., Mn Ž1230., Zn Ž59., Sr Ž39., Co Ž13., Ni Ž12., Cu Ž2. and less than 1 kg of Cd᎐Cr᎐As᎐Pb. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Estuarine sediments; Metal leaching; Dredging; Sulfur oxidation
U
Corresponding author. Tel.: q358-50-5740494; fax: q358-2-2154818. E-mail address: [email protected] ŽP. Peltola..
0048-9697r02r$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 0 1 . 0 0 8 7 2 - 5
110
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
1. Introduction In coastal areas of Finland, a large number of streams are periodically severely acidified ŽpH down to 2.8. and are loaded with toxic concentrations of metals such as Al Žup to 250 mgrl., Mn Žup to 15 mgrl., Ni Žup to 1 mgrl. and Cd Žup to ˚ ¨ and Bjorklund, 10 grl. ŽWeppling, 1993; Astrom ¨ ˚ ¨ and Spiro, 2000.. The reason for 1995; Astrom the low pH values and the high metal loads in these streams is not point source pollution, but primarily, an extensive leaching of sulfide-rich marine sediments occurring abundantly in the ˚ ¨ and Bjorklund, coastal areas ŽAstrom 1996; ¨ ˚ ¨ and Astrom, ˚ ¨ 1997; Eden Astrom ´ et al., 1999.. In the natural state, these sediments do not release acids and metals, but when artificially drained for agricultural purposes, which has been done for an area of at least 3000 km2 along the coast, they oxidise and transfer part of their high natural content of potential acidity and heavy metals into ˚ ¨ 1998.. As a consedrains ŽPalko, 1994; Astrom, quence, stream-water pollution with associated hydroecological damage are frequent phenomena along the coast, in particular in areas of intense agricultural activity Že.g. Kjellman and Hudd, 1996.. In the flat terrain in the coastal areas of Finland, a variety of factors including natural phenomena Žisostatic land uplift. and human activities result in a need to dredge the middle and lower reaches as well as estuaries of many streams and rivers. This dredged material, which is rich in sulfide minerals and leachable heavy metals, is often deposited on the banks and, thus poses a potential threat to the aquatic environment. Hence, in addition to the extensive ditching of farmland, dredged masses most likely contribute to the acidification and metal loading of the surface waters in the coastal areas. The main aim of this study was to characterise, prior to the initiation of extensive dredging operations at many coastal sites, the geochemistry of lake sediments, which in the near future is likely to be dredged. The study consists of determination of concentrations, associations and spatial distribution of a variety of metals and non-metals
in such sediments, and an air-oxidation experiment in laboratory, by which the extent of acidity production and metal release on oxidation of the sediments can be assessed. We selected the southern part of the Larsmo lake Žmid-western Finland., which is an interesting area as it is strongly affected by isostatic land uplift Žapprox. 8 mmra. and was converted from a brackish-water bay to an artificial fresh-water lake in 1962.
2. Area description The data presented under this heading is mostly taken from Kalliolinna Ž1996.. The total drainage ¨ lakes is 4280 km2 area of the Larsmo and Oja ŽFig. 1.. The total area of the lakes at the fixed water level Ž N60 " 0.00 m. is 85 km2 . The area of ¨ lake 12 km2 . The Larsmo lake is 73 km2 and Oja total water volume is 200 million m3. The average ¨ water depth in Larsmo lake is 2.6 m and in Oja lake 1.6 m. Maximum water depths are less than 15 m. Four rivers discharge into Larsmo lake. Of these, the Esse river is largest Ždrainage area s 2054 km2 , median water flow 1989᎐1995 s 15.8 m3rs, account for 45% of the total inflow to the lake., followed by Purmo river Ž864 km2 , 6.9 m3rs, 25%., Kronoby river Ž788 km2 , 6.1 m3rs, 22%. and Kovjoki river Ž292 km2 , 2.4 m3rs, ; 10%.. The Esse and Purmo rivers have a joint mouth close to the lake ŽFig. 1.. The rivers are characterised by occasionally acidic waters rich in humic substances, Fe, N and P. The buffering capacity of the waters is low and occasionally almost absent. The Larsmo lake has had several mass kills of fish, which have been explained by ditching of farmlandrforests in the catchment and also to dredging of the river mouths. The fish mass kills have occurred mainly during spring and autumn high water flows. The local industries use a maximum of 5 m3rs of the fresh water from Larsmo lake and 1 m3rs ¨ lake. The remaining water in the lakes from Oja Žminus that volume which is evaporated. eventually discharges into the Botnian sea via locks at Hastgrundet and Gertruds ŽFig. 1.. The average ¨ outflow from Larsmo lake in 1989᎐1995 was 33
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
111
¨ lake. The sampling area Ža. and a dredging area Žb. are indicated. Fig. 1. Regional map over Larsmo-Oja
m3rs, of which 26 m3 Ž78%. went through Hastgrundet. The flow during this period ranged ¨ ¨ from 20.7 to 47.2 m3rs. The outflow points of Oja lake are at the Krakila ¨ ¨ and Palma locks ŽFig. 1.. These locks are, however, being used only at exceptionally high water levels. The outflow from ¨ ja lake is only 1% of that from Larsmo lake. O The annual amounts of material deposited in the lakes are as follows Ždata from 1989᎐1995.: 4100 t solids; 150 t nitrogen; and 14 t phosphorus. The annual material flow to the sea during this period was 6500 t solids, 940 t nitrogen and 48 t phosphorus. Of these amounts 30% comes from the Esse and Purmo rivers, 25% from the Kronoby river and 10% from the Kovjoki river. The material transported to the sea flows almost en-
tirely through Larsmo lake and only to some ¨ lake Ž1᎐2%.. extent via Oja
3. Methods 3.1. Sediment sampling Sampling of sediments at 39 sites was carried out during the summer of 1998 in the southernmost part of the Larsmo lake at the mouths of Esse, Purmo and Kovjoki rivers ŽFig. 1, area a.. The overall aim of the sampling was to collect the uppermost black sediment, which is present throughout most of the lake to depths of more than 2 m below the water-sediment interface.
112
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
These sediments consists of silt Žunpublished data of similar material from the same area.. The sample tool was a peat drill ŽRussian sampler., through which an undisturbed semicircular sample measuring 50 cm in length and 4.6 cm in radius is collected. A surface sample was taken from 0 to 50 cm and a deep sample from 50 to 100 cm. At three sampling points less than 1 m of the black mud was present and only the upper sample was taken, giving a total of 75 samples. The samples were immediately frozen in order to prevent oxidation. At the sampling points, the water depth varied between 0.2 and 2.5 m. All samples were black coloured with occasional occurrences of thin grey stripes, except two samples at the Kovjoki river mouth, which had a brown colour towards the surface. 3.2. Sediment analysis A portion of each sediment sample was oven dried below 60⬚C and comminuted in an agate mortar. A 2.0-g portion of the pulverised material was analysed for Al Ž100., As Ž2., B Ž3., Ba Ž1., Ca Ž100., Cd Ž0.01., Co Ž1., Cr Ž1., Cu Ž1., Fe Ž100., K Ž100., Mg Ž100., Mn Ž2., Mo Ž1., Na Ž100., Ni Ž1., P Ž10., Pb Ž2., Sr Ž1., V Ž1. and Zn Ž1. with ICP-AES after an aqua regia Ž95⬚C 2:2:2 HClrHNO3rH 2 O. digestion of the sample Žthe numbers in parenthesis are the detection limits in ppm.. This analysis was carried out in a commercial accredited ŽISO 9002. geochemical laboratory. In two samples Ž2r75. boron was below the detection limit. All other elements in all sediment samples were above the detection limits. Sulfur and carbon were determined in the same laboratory with Leco. The analyses of duplicates of 13 random samples Žnine for C and S. showed that for each element the data variance is significantly Ž99% confidence level. larger than the error variance. Aqua regia is a strong partial leach that mainly digests trioctahedral micas, clay minerals, salts and secondary precipitates ŽRaisanen, 1999.. ¨ ¨ 3.3. Sediment oxidation experiment and water analysis The oxidation experiment, was done under con-
trolled laboratory conditions and was designed to simulate sulfur oxidation and elemental leaching as they occur under field conditions. The experiment is a modification of a similar experiment ˚ ¨ and Bjorklund Ž1997.. For each done by Astrom ¨ sediment sample a portion of 3.0 g Ždry wt.. of reduced sediment was placed in a plastic tube and mixed with 9 ml of deionised water. The pH and conductivity were measured from the clear supernatants Žreducing conditions.. The samples were then kept at room temperature for 1 year and wetted with 3 ml of deionised water every month. After this period 9 ml of water was added and the tubes were shaken. After the sediment particles had settled, the pH and conductivity of the clear supernatants were measured Žoxidised state.. Control measurements were performed on the following day. In this experiment no leaching was allowed, since the added water Ž9 ml plus several 3-ml portions. was left to evaporate leaving the concentration of elements intact. After the measurements of the samples in the oxidised state, the supernatants were analysed for Al Ž1., As Ž1., Cd Ž0.02., Co Ž0.02., Cr Ž0.1., Cu Ž0.1., Li Ž0.5., Mn Ž0.05., Na Ž20., Ni Ž0.2., Pb Ž0.5., Rb Ž0.01., Sr Ž0.01., U Ž0.05. and Zn Ž0.5. with ICP-MS Žthe numbers in parentheses are the detection limits in ppb.. The analytical precision of the ICP-MS analysis was determined with in-house standard materials, reagent blanks and five replicates. The analytical precision was satisfactory and the ratio of data to error variance high for the determined elements. Chromium and As had 5r75 and 3r75 respectively, samples below detection. Some of the sediment samples were accidentally Žpartly. oxidised in the laboratory prior to the first measurement and their pH and conductivity values in reduced states was therefore not measured. These samples were not included in the statistical analyses. In the laboratory experiment, the waterrsoil ratio Žweight., with and without the monthly additions, were 13.3 and 3.0, respectively. In the field, the waterrsoil ratio for a period of 1 year Ž500 mm precipitation. is 1.78᎐2.23 and only 0.45 when the pore water has dried. Therefore, in the oxidation experiment there was excess water as com-
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
pared to the natural conditions. The temperature in the laboratory is approximately 20⬚C, while in the study area it is above 15⬚C on average for only 3 months per year ŽAtlas of Finland, 1987.. Hence, the results of the experiment most likely correspond to several years of oxidation under natural field conditions. The rate of oxidation in-field is of course affected also by numerous other factors such as the sulfur-bacterial activity, the thickness Žair penetration. of the sediments and climatic variables. 3.4. Statistical analyses The statistical analyses were performed with robust and non-parametric methods Žrange, median, Spearman correlation, Mann᎐Whitney Utest., which means that there is no requirement of normal data distributions. The data was examined for normal distribution with the Kolmogorov᎐ Smirnov test and with histograms. Approximately one half of the elements had a normal distribution Ž95% level., which motivates the use of nonparametric statistical methods. In most geochemical data analysis, non-parametric methods are preferred over those which require a normal data distribution ŽReimann and Filzmoser, 1999.. Growing symbol maps were prepared in order to graphically show the spatial variation in the data and to simplify the interpretation of possible geochemical distribution patterns.
4. Results and discussion 4.1. Element concentrations and associations The aqua regia extractable concentrations of elements in the Larsmo lake sediments are similar to those in 317 samples of similar sediments ˚ ¨ and Bjorklund, ŽAstrom 1997. distributed on the ¨ coastal plains of central western Finland ŽTable 1.. The relatively small, but significant, geochemical differences that exist between these two data sets are most likely related to different depositional environments, since while the sediments col-
113
lected from the coastal plains represent a heterogeneous group with an unknown mixture of lacustrine and marine sediments of varying age, the studied ŽLarsmo lake. sediments have been deposited within a geographically restricted area and are thus formed in a specific environment. In the Larsmo lake sediments, the concentrations of Na, B and S are significantly higher and Zn significantly lower in the deep Ž50᎐100 cm. than in the surface Ž0᎐50 cm. layer. The other elements do not occur in significantly different concentrations in the two layers ŽMann᎐Whitney U-test, 95% level.. The enrichment of Na and B in the deep layer is explained by a higher proportion of marinerbrackish Žcontains trapped NaCl and borate. to fresh water sediments in this than in the surface layer. This argument is in agreement with Eden ´ Ž1994., who found that the Na concentration in overbank sediments Ždeposited under fresh-water conditions. along 49 streams throughout Fennoscandia did not exceed 0.06%, a figure which is clearly lower than those in the deep layer of the Larsmo lake sediments Ž0.04᎐0.31%.. The ‘depletion’ of S in the surface layer is probably due to less reducing conditions here than deeper down in the sediment column. The maximum concentrations of the potentially toxic elements As, Ba, Cd, Co, Cr, Cu, Mo, Ni, Pb, V and Zn in the sediment samples were lower than the limit values for contaminated soils by factors of 3᎐100 ŽTable 1.. The only element that exceeded the limit value was S ŽTable 1.. The high S concentration is, however, considered normal since reducing conditions in the sediments have resulted in sulfide formation from sea-water sulfates and from sulfate derived from the extensively ditched S-rich fine-grained sediments in the drainage area ŽPalko et al., 1986.. High sulfur concentration is a typical feature also in many other reducing marine sedimenters elsewhere in the Botnian sea ŽMuller, 1999.. ¨ In the deep sediment layer, the elements can be divided into two main groups according to the interelement correlations and scattergram features: Ž1. Ca, K, Mg, Sr, B, Na, V, Cr, Ni, S and Cu; Ž2. Fe, Mn, Mo, Zn, Cd, Co, Ba, As, P, Pb and C. The highest coefficient for group 1 is
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
114
Table 1 Aqua regia extractable element concentrations in sediments in the southern Larsmo lake Žsurface and deep layer. and in ˚ ¨ 1998.. The last column shows the recommended corresponding sediments on the coastal plains ŽCP. of Finland Ž n s 317. ŽAstrom, and limit values for potentially toxic elements in soilsrsediments ŽAssmuth, 1997. Surface layer Ž0᎐50 cm.
Fe Ž%. C Ž%. Al Ž%. S Ž%. Mg Ž%. Ca Ž%. K Ž%. P Ž%. Na Ž%. Mn Žppm. Zn Žppm. Ba Žppm. V Žppm. Sr Žppm. Cr Žppm. Ni Žppm. Cu Žppm. Co Žppm. As Žppm. Pb Žppm. B Žppm. Mo Žppm. Cd Žppm.
Med
Max
Min
Med
Max
CP Ž100᎐350 cm. Med
Rec.rLim.
Min
Deep layer Ž50᎐100 cm.
1.46 1.18 0.56 0.16 0.25 0.27 0.12 0.08 0.02 220 35 26 19 15 13 8 7 6 4 5 bd 1 0.1
4.29 2.43 1.19 0.58 0.50 0.45 0.24 0.13 0.08 527 75 55 38 35 29 17 13 13 9 8 6 3 0.2
5.91 5.33 1.43 1.97 0.59 0.55 0.29 0.17 0.30 875 89 69 45 53 43 20 20 17 17 14 13 4 0.3
2.12 1.24 0.61 0.19 0.30 0.30 0.14 0.08 0.04 214 37 27 21 18 16 8 9 6 4 5 3 2 0.1
3.96 2.22 1.14 0.73 0.53 0.47 0.26 0.13 0.13 563 62 51 39 40 29 17 15 11 10 7 11 3 0.2
5.89 6.30 1.30 1.33 0.62 0.71 0.31 0.17 0.31 1176 93 66 46 66 55 19 22 16 20 16 15 4 0.3
3.80 1.37 2.02 0.54 1.04 0.50 0.60 0.06 0.10 448 90 87 50 36 48 31 27 13 ᎐ 12 ᎐ ᎐ ᎐
᎐ ᎐
recorded between Ca and Sr Ž rs s 0.93. and for group 2 between Co and Zn Ž rs s 0.95.. The highest coefficient recorded between one element from each group is that between V and Zn Ž rs s 0.40.. Aluminium has positive correlations with elements from both of the groups, but overall, higher with those of group 2. In the surface layer the correlations are generally similar except for the correlations involving elements which occur in considerably different concentrations ŽS, B and Na. in the two layers. For example, in the deep layer the Spearman correlation between Mn and S is y0.21 and in the upper layer 0.57, while for Fe and S the corresponding correlations are 0.0 and 0.55, respectively. The spatial geochemical patterns in the sediments in the southern Larsmo lake are affected by the general geochemical behaviour of the elements, the settling rate of different organic and
2 = 10y5 ᎐ ᎐ ᎐ ᎐ ᎐ ᎐ 90r700 600r600 50r500 ᎐ 80r500 40r300 32r400 50r200 13r60 38r300 ᎐ 5r200 0.5r10
inorganic colloidsrparticulates and by sediment resuspension and redeposition. Since the water depth in the study area is low Ž- 2.5 m. resuspension of the sediments may occur due to water currents, ice-cover movement and the frequent motor-boat traffic in summer. The spatial variations in the S concentrations in both sediment layers are shown in Fig. 2. Because the concentrations of the other elements were relatively low Žfar below limit values. and did not show any distinct spatial patterns, they are not shown. 4.2. Electric conducti¨ ity and pH The median pH redrox in the samples of the deep layer is 7.1r3.4 and the median conductivityredrox is 1055r3916 Srcm. In the surface layer the corresponding figures are 7.0r3.4 and 585r3035 Srcm, respectively. As these me-
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
115
Fig. 2. The distribution of sulfur in the sediments in the southern Larsmo lake Žthree dots are missing in the deep layer..
dian values and Fig. 3 show, the 1 year oxidation resulted in a substantial lowering of the pH values in all samples. This is caused by the oxidation of reduced sulfur by atmospheric oxygen, resulting in an extensive production of acids Žsulfuric acid and Fe. and thus a lowering of pH as the buffering capacity of the sediments is low Žlow contents of carbonate.. Higher sulfur concentrations generally produced lower pH ox , as is clearly shown by the correlations between these variables Ž rs s y0.50 and y0.56. ŽTable 2.. This relationship is further indicated by the fact that in 51% of the samples with the bottom 50% of the S concentrations the pH ox - 3.5, while for the samples with the top 50% of the S concentrations the corresponding figure was 71%. The concomitant strong increase in the median conductivity values Ž519% increase for the surface-layer samples and 372% increase for the deep-layer samples. shows that ions are formed and mobilised in large
amounts by the oxidation and acidification of the sediments ŽFig. 4.. The conductivity in the reduced state is controlled by the extent of trapped sea salts. This explains the higher conductivityred values of the deep-layer sediments, which in contrast to the surface sediments, to a large extent have been deposited in brackish Žand not fresh. water. 4.3. Mechanisms of metal release on oxidation The greatest relative leachability Žoxidation experiment. in the two layers are shown by Na Ž9.9᎐12.3%., Mn Ž9.6᎐9.7%., Cd Ž8.7᎐9.5%. and Co Ž4.3᎐5.2%. ŽTable 3.. Cobalt, Cd and Zn show an increase in the leached amounts for the surface layer ŽFig. 5.. This indicates that these metals are more weakly bound to the surface sediments, probably caused by antropogenic input in
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
116
Fig. 3. pH of the sediment samples in reduced and oxidised state. The loose dots represent the samples, which were accidentally exposed to air prior to the first measurements. The pH of these samples in the reduced state was therefore not measured.
recent decades. Another possible cause to the higher leachable fractions in the surface layer is the change from brackish to fresh water conditions 40 years ago. This decrease in salinity in the depositional environment during the last few decades explains the lower amounts of leachable Na in the surface layer.
The correlations between pH and the concentrations of Al, As, Cr, Cu, Li, Ni and U in the Žoxidised. leachates were inverse and high Žat least in one layer above 0.70, Table 4.. The critical pH values, which were obtained graphically from scattergrams and which indicate the pH point where a strong increase in the elemental
Table 2 Spearman correlation coefficients between the amounts of elements mobilised and the pH and conductivity developed by the oxidation of lake sediments Žfirst column., and the total S, total C and aqua regia extractable element concentrations in the sediments Sulfur
Carbon
Surface Na Al Mn Zn Sr Co Ni Cu Cd Cr Pb As Li Rb U pHox Condox U
U
0.49 0.56U 0.38U 0.42U 0.21 0.21 0.63U 0.70U 0.31 0.41U y0.18 0.36U 0.66U y0.44U 0.72U y0.50U 0.75U
Significant at the 95% level.
Deep y0.08 0.49U y0.25 y0.02 y0.05 0.20 0.61U 0.54U 0.01 0.38U 0.01 0.31 0.60U y0.48U 0.52U y0.56U 0.42U
Surface y0.24 y0.43U y0.22 y0.26 y0.40U y0.06 y0.58U y0.66U y0.35U y0.17 0.00 y0.45U y0.54U 0.19 y0.67U 0.43U y0.54U
Corresp. metal Deep 0.20 y0.29 0.10 y0.15 y0.05 0.04 y0.67U y0.66U y0.28 y0.33U y0.21 y0.44U y0.55U 0.54U y0.58U 0.42U y0.52U
Surface U
0.53 y0.20 0.91U 0.47U 0.46U 0.64U 0.52U 0.31 0.36U 0.06 0.02 0.31 ᎐ ᎐ ᎐ ᎐ ᎐
Deep 0.80U y0.23 0.87U 0.35U 0.63U 0.37U 0.30 0.29 0.18 y0.01 0.07 y0.12 ᎐ ᎐ ᎐ ᎐ ᎐
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
117
Fig. 4. Electric conductivity of the sediment samples in reduced and oxidised state. The loose dots represent the samples, which were accidentally exposed to air prior to the first measurements. The electric conductivity of these samples in the reduced state was therefore not measured.
concentrations begins were as follows: Ni 4.8; Cd 4.4; Co 4.0; Li 4.0; Al 3.9; As 3.9; Cu 3.8; U 3.8; and Cr 3.3 ŽFig. 5.. Zinc correlated relatively weakly with pH ŽTable 4. but the scattergram gives a critical pH value of approximately 4.8. Note, however, that in these results the pH spec-
trum is between 4.8 and 3.0, which means that the study of the elemental behaviour is limited to this pH range. Among the elements that correlated with the pH, three different patterns in the concentrationrpH scattergrams were observed. Ž1. The profiles for Al, ŽAs., Li, Cu and U are char-
Table 3 Easily mobilisable amounts Žppm. of metals and As Žas determined by the oxidation experiment. in two layers of lakermarine sediments. The last column Ž%. shows the relative amounts of easily mobilised elements as compared to the aqua regia extractable concentrations Surface layer Ž0᎐50 cm.
Na Al Mn Zn Sr Co Ni Cu Cd Cr Pb As Li Rb U
Deep layer Ž50᎐100 cm.
Min
Med
Max
%
Min
Med
Max
%
8.7 1.1 6.7 - 0.01 0.20 0.02 0.04 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01
83 71 55 2.7 1.4 0.65 0.47 0.06 0.02 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01
280 222 107 6.1 3.2 1.3 1.0 0.28 0.03 0.10 0.03 0.01 - 0.01 - 0.01 - 0.01
9.9 0.66 9.6 3.8 4.5 5.2 3.2 0.59 9.5 0.04 0.10 0.03 ᎐ ᎐ ᎐
25 0.94 2.7 0.14 0.30 0.04 0.08 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01
174 68 53 2.2 1.6 0.50 0.46 0.06 0.01 0.01 0.01 - 0.01 - 0.01 - 0.01 - 0.01
401 156 139 4.9 3.6 0.80 0.87 0.16 0.02 0.05 0.03 0.01 - 0.01 - 0.01 - 0.01
12.3 0.6 9.7 3.5 4.3 4.3 2.7 0.48 8.7 0.04 0.09 0.24 ᎐ ᎐ ᎐
118
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
Fig. 5. The relationship between pH Žand aqua regia extractable Mn concentrations . and selected element concentrations in the aquatic phase that had equilibrated with sediment samples oxidised for a period of 1 year in the laboratory.
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
119
Fig. 5. Ž Continued..
acterised by low concentrations at high pH and rapidly increasing concentrations at a given pH value, thus making it easy to obtain critical pH values ŽFig. 5.. Ž2. The concentrations of Cd, Co, Ni and Zn increase in a linear manner with decreasing pH values ŽFig. 5.. From these patterns it is more difficult to estimate a critical pH value since it probably lies at a higher pH than 5.0. Ž3. Chromium is the most immobile of the ‘pH sensitive elements’, showing a slight concentration increase only at very low pH values ŽFig. 5.. The elements whose mobility were least or not at all affected by the decreasing pH, in the observed pH range, were Na, Mn ŽFig. 5., Pb, Sr
and Rb. The concentration of Rb, in contrast to all other studied elements, increased in the solution phase as the pH of this became higher ŽTable 4.. Aluminium is derived from weathered silicates and is probably immobile as hydroxides and oxides at more neutral pH conditions. Al-hydroxysulfates in soils dissolve at pH - 4.2 ŽReimann and de Caritat, 1998.. Also Cr, Li and U have probably been bound as hydroxides and oxides. The elements As, Cd, Co, Ni and Zn probably derive from metal sulfides, which are stable at neutral and reducing conditions but which dissolves Žand oxidise. under aerobic conditions, to
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
120
Table 4 Spearman correlation coefficients between the amounts of elements mobilised and the pH and conductivity developed by the oxidation of sediments for 1 year pH
Na Al Mn Zn Sr Co Ni Cu Cd Cr Pb As Li Rb U U
Conductivity
Surface
Deep
Surface
Deep
0.23 y0.95U y0.24 y0.38U 0.01 y0.36U y0.65U y0.83U y0.31 y0.58U 0.38U y0.48U y0.8U 0.50U y0.79U
0.10 y0.93U 0.27 y0.42U 0.22 y0.54U y0.80U y0.89U y0.45U y0.76U y0.07 y0.80U y0.96U 0.73U y0.86U
0.49U 0.62U 0.60U 0.62U 0.50U 0.39U 0.76U 0.76U 0.58U 0.26U y0.28 0.48U 0.79U y0.57U 0.82U
0.68U 0.52U 0.12 0.36U 0.35U 0.20 0.65U 0.67U 0.33 0.34U y0.01 0.53U 0.64U y0.56U 0.67U
Significant at the 95% level.
form a metal Žor As. ion and sulfate. The results by Eden ´ Ž1994. indicate that a Na concentration above 0.06% in sediments is derived from trapped sea salts ŽNaCl.. The fact that the deep layer has higher easily leachable concentrations of Na than this means that the sediments have, to a greater extent, been deposited under brackish conditions. The leachable fractions of Mn Žhydroxides and oxides. probably dissolve at an earlier stage in the oxidation process, at pH values somewhere between 5 and 7. The soluble Mn fraction shows a strong correlation with the aqua regia extractable Mn concentration ŽTable 2, Fig. 5., which means that the proportion of easily soluble species Žrelative to aqua regia extractable Mn., such as Mn oxyhydroxides is constant throughout the sediments and controls the amounts of Mn leached. The pattern for Sr is somewhat similar to that of Mn, but the Spearman correlation for the leachable and aqua regia extractable Sr fraction is weaker than the corresponding one for Mn ŽTable 2.. Lead has probably been bound as sulfide in the sediments. Its leaching proportion is, however, relatively low ŽTable 3. which may be explained by sorption onto soilrsediment constituents after its release from the sulfides.
4.4. An assessment of en¨ ironmental impacts of dredging Since the isostatic land uplifting will continue for several thousand years in the coastal regions of central Finland and Sweden, dredging operations will certainly be frequent in these regions in the future, mainly in order to: Ž1. clear out sediments that accumulate at the river mouths, in which otherwise waterflow velocity will decrease with a concomitant risk of increased flooding; Ž2. deepen the sailing depths close to harbours; and Ž3. improve the recreational values along the shores of lakes and the coast. At present, a number of large dredging operations are planned in a variety of estuarinerlacustrine settings along the coasts. One of these is located east of the town of Jakobstad Žclose to the study area, Fig. 1, area b., where an area of 335 800 m2 of a shallow lake is to be dredged in the near future. In this area approximately 50 000 m3 of sediment will be dredged from the uppermost 1 m of the sediment column. The material to be dredged equals 23 300 t of sediment Ždry wt.. with a volume of 10 500 m3. Preliminary Žunpublished. analyses have shown that these sediments are geochemically similar to those of this study, except that they have even higher S concentrations Žmean s 1.0%, mean in this study s 0.71%. and thus probably consist an even larger environmental threat than the studied sediments. According to the median values Žppm. of the elements leached in the oxidation experiment ŽTable 3., these masses has the potential to release Žin the short term. considerable amounts Žkg. of Al Ž1710., Mn Ž1230., Zn Ž59., Sr Ž39., Co Ž13., Ni Ž12., Cu Ž2. and Cd᎐Cr᎐As᎐Pb Žless than 1 kg of each.. Because the dredged material will be deposited close to the site where it was collected, the mobilised elements could to a large extent end up in the lake, which is vulnerable to overloading due to its small water volume and low water circulation rates. While being somewhat speculative due to the large differences that exist between the conditions in field and laboratory, these figures highlight the fact that not only the potential acidity, but also the metals within the dredged masses pose a potential threat to the local aquatic environment. While these sediments
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
will be limed ŽCaCO 3 . and their pH thus kept relatively high at least in the short term, it is questionable whether such procedures will be enough to prevent acid and metal leakage in the long term because, Ž1. thorough liming of dense sediment masses is a difficult task, Ž2. if at some stage the acidity within the sediment becomes greater than the buffering capacity of the added lime, many ‘pH sensitive’ metals Že.g. Al, As, Cd, Co, Cu, Li, Ni, U. are likely to rapidly mobilise at decreasing pH ŽTable 4, Fig. 5. and Ž3. at the examined pH values ŽpH 4.8᎐3.0. some of the metals ŽMn and Sr. in the sediments are controlled by factors other than pH ŽFig. 5. and could thus become mobile even if the liming measures are successful. These figures and arguments thus suggest, that the potential acidity as well as metals within the sediments should be considered when measures are taken to minimise the environmental impact of the planned dredging operations.
5. Conclusions The concentrations of toxic elements in the Larsmo lake sediments are low and are as such no threat to the environment. However, the sediments contain large amounts of reduced sulfur, which is normal and natural, but which result in extensive acidity production and concomitant metal mobilisation on oxidation of the sediments. While under natural conditions, oxidation of the sediments occur at a low rate when the isostatic land uplifting exposes sediment layers to atmospheric oxygen, man made operations such as ditching and dredging quickly exposes large volumes of reduced sediments for the atmospheric oxidant. In the study area, the harmful effects of sediment oxidation is expected to be exceptionally high because of poorly buffered and biologically sensitive surface waters. In the highlighted dredging operation of 50 000 m3 of sediment, enough metals can be released to substantially damage the local aquatic environment. Consequently, in large dredging operations such as this one, care must be taken when selecting a site for dumping of the masses. Dumping close to
121
areas with high water circulation is recommended. If possible, the dredged material should be dumped as sub-surface fillings, which would prevent the oxidation of the reduced S. Dumping of dredged masses close to marine environments is preferred since marine water has far greater buffering capacity than fresh water. Neutralisation of dredged material is of great importance and is not to be neglected. While the oxidation experiment worked very well, it could still be improved to give more detailed results. For example, to allow for stepwise leaching where water is analysed continuously to get a more detailed pattern of acidity and metal mobilisation. The waterrsoil ratio could be optimised to be more close to that in field. To simulate the local climate may be unnecessary since the experiment would then be too time consuming. The sampling procedure was considered adequate and detailed enough for the purposes of the study. The differences in the Na concentrations in the layers highlighted the fact that Na Žand Cly. concentration can be used as a dating agent in this type of environment, if samples are collected at a much higher vertical frequency than was done in the present study. Acknowledgements The authors would like to thank for the financial support provided by the city of Jakobstad, UPM-Kymmene ŽJakobstad mills., Finska Vetenskaps-societenrSohlbergska delegationen and the Waldemar von Frenckel grant. The authors are also grateful to Soren Frojdo ¨ ¨ ¨ for help with the ICP-MS analyses and the figures, and to Peter ¨ sterholm for valuable advises during the project. O References ˚ ¨ M, Astrom ˚ ¨ J. Geochemistry of stream water in a Astrom catchment in Finland affected by sulfidic fine sediments. Applied Geochemistry 1997;12:593᎐605. ˚ strom A A. Impact of acid sulfate soils on stream ¨ M, Bjorklund ¨ water geochemistry in western Finland. J Geochem Explor 1995;55:163᎐170. ˚ strom A A. Geochemistry and acidity of ¨ M, Bjorklund ¨ sulfide-bearing postglacial sediments of western Finland. Environ Geochem Health 1997;19:155᎐164.
122
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
˚ strom A A. Hydrogeochemistry of a stream ¨ M, Bjorklund ¨ draining sulfide-bearing postglacial sediments in Finland. Water Air Soil Pollut 1996;89:233᎐246. ˚ strom A ¨ M. Mobility of Al, Co, Cr, Cu, Fe, Mn, Ni and V in sulfide-bearing sediments exposed to atmospheric O 2 : an experimental study. Environ Geol 1998;36:219᎐226. ˚ strom A ¨ M, Spiro B. Impact of isostatic uplift and ditching of sulfidic sediments on the hydrochemistry of major and trace elements and sulfur isotope ratios in streams, western Finland. Environ Sci Technol 2000;34:1182᎐1188. Atlas of Finland 1987. Climate folio 131. National Board of Survey and Geographical Society of Finland, 1987, pp. 32. Assmuth T. Analysis and proposals of guideline values for concentrations of hazardous substances in soil Žin Finnish.. Suomen Ymparisto ¨ ¨ Keskuksen Moniste 1997;92:56. Eden ´ P. Wide-spaced sampling of overbank sediment, till, humus and river water in Fennoscandia, PhD Thesis, Abo Academy University, 1994. Eden ´ P, Weppling K, Jokela S. Natural and land-use induced load of acidity, metals, humus and suspended matter in Lestijoki, a river in western Finland. Boreal Environ Res 1999;4:31᎐43. ¨ ¨ Kalliolinna M. Luodon-Ojanjarven vedenlaatu vuosina 1989᎐1995. The association of water protection in western Finland Žin Finnish., 1996, p. 52. Kjellman J, Hudd R. Changed length-at-age of burbot, Lola
lota, from an acidified estuary in the Gulf of Bothnia. Environ Biol Fish 1996;45:65᎐73. Muller A. Distribution of heavy metals in recent sediments in ¨ the Archipelago Sea of southwestern Finland. Boreal Environ Res 1999;4:319᎐330. ¨ ¨ Palko J., Rasanen M. and Alasaarela E. Luodon-Ojanjarven ¨¨ valuma-alueen maaperan ¨ ja vesiston ¨ happamuuskartoitus. Vesi ja ymparistohallituksen julkaisuja 1986, p. 11 Žin Fin¨ ¨ nish.. Palko J. Acid sulfate soils and their agricultural and environmental problems in Finland. Acta Univ. Oul. C 75, PhD Thesis, University of Oulu, 1994. Reimann C, de Caritat P. Chemical elements in the environment. Berlin Heidelberg New York: Springer-Verlag, 1998:398. Reimann C, Filzmoser P. Normal and lognormanl data distribution in geochemistry: death of a myth. Consequences for the statistical treatment of geochemical and environmental data. Environ Geol 1999;9Ž39.:1001᎐1014. Raisanen M-L. The podzolisation in Finland, processes and ¨ ¨ methods of analysis. Conference of Environmental Geology 23᎐24 March 1999, TurkurFinland Žin Finnish., 1999, pp. 39᎐43. Weppling K. Hydrochemical factors affecting the neutralization demand in acid sulfate waters. Vatten 1993;49:161᎐170.
Concentrations and leachability of chemical elements in estuarine sulfur-rich sediments, W. Finland ˚ ¨ Pasi PeltolaU , Mats Astrom ˚ Akademi Uni¨ ersity, Abo ˚ 20500, Finland Department of Geology, Abo Received 1 February 2001; accepted 9 April 2001
Abstract Concentrations, distributions and mobility of chemical elements were investigated in reduced sulfur-rich estuarine sediments located in western Finland. The main objective was to determine the possible extent of metal leaching when dredged masses of these sulfur-rich sediments are dumped on the land and thus exposed to air. When dredged, the reduced sulfur in the sediments oxidises resulting in a lowering of pH, which in turn is expected to leach metals. The study area is an artificial lake claimed from the Botnian sea in 1962. In this lake, several mass-kills of fish have occurred, believed partly to be due to dredging. Two sediment samples Ž0᎐50 and 50᎐100 cm. were taken from 39 sampling points in the lake. These samples were leached in aqua regia Ž2:2:2 HNO3rHClrH 2 O. and analysed for Fe, Al, Mg, Ca, K, P, Na, Mn, Zn, Ba, V, Sr, Cr, Ni, Cu, Co, As, Pb, B, Mo and Cd with ICP-AES. Sulfur and organic carbon were analysed with Leco. In a controlled laboratory experiment, the sediments were allowed to oxidise for 1 year while moisturised with deionised water every month. The pH and conductivity were determined in the beginning of the experiment Žreduced state. and in the end Žoxidised state.. In the supernatants in the oxidised states the amount of leached metals ŽNa, Al, Mn, Zn, Sr, Co, Ni, Cu, Cd, Cr, Pb, U, Li, Rb and As. were determined with ICP-MS. The sediments were found to contain low levels of toxic metals but, as expected, high concentrations of sulfur. In the experiment, pH was lowered Ždown to 3.0. and the conductivity increased in all samples due to oxidation and release of metal ions. The extent of leaching varied between 0.03% for As and 12.3% for Na. Critical pH values, at which high amounts of metals begin to leach, were obtained graphically. These values varied between 4.8 ŽNi. and 3.3 ŽCr.. Not all elements were controlled by pH, e.g. Mn correlated well with its aqua regia leachable concentration. In a planned dredging operation in the area some 23 300 t Ž10 500 m3 . Ždry wt.. of sediments will be dredged. The amounts of metals likely to be leached, according to the results from this study, are as follows Žkg.: Al Ž1710., Mn Ž1230., Zn Ž59., Sr Ž39., Co Ž13., Ni Ž12., Cu Ž2. and less than 1 kg of Cd᎐Cr᎐As᎐Pb. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: Estuarine sediments; Metal leaching; Dredging; Sulfur oxidation
U
Corresponding author. Tel.: q358-50-5740494; fax: q358-2-2154818. E-mail address: [email protected] ŽP. Peltola..
0048-9697r02r$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 0 1 . 0 0 8 7 2 - 5
110
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
1. Introduction In coastal areas of Finland, a large number of streams are periodically severely acidified ŽpH down to 2.8. and are loaded with toxic concentrations of metals such as Al Žup to 250 mgrl., Mn Žup to 15 mgrl., Ni Žup to 1 mgrl. and Cd Žup to ˚ ¨ and Bjorklund, 10 grl. ŽWeppling, 1993; Astrom ¨ ˚ ¨ and Spiro, 2000.. The reason for 1995; Astrom the low pH values and the high metal loads in these streams is not point source pollution, but primarily, an extensive leaching of sulfide-rich marine sediments occurring abundantly in the ˚ ¨ and Bjorklund, coastal areas ŽAstrom 1996; ¨ ˚ ¨ and Astrom, ˚ ¨ 1997; Eden Astrom ´ et al., 1999.. In the natural state, these sediments do not release acids and metals, but when artificially drained for agricultural purposes, which has been done for an area of at least 3000 km2 along the coast, they oxidise and transfer part of their high natural content of potential acidity and heavy metals into ˚ ¨ 1998.. As a consedrains ŽPalko, 1994; Astrom, quence, stream-water pollution with associated hydroecological damage are frequent phenomena along the coast, in particular in areas of intense agricultural activity Že.g. Kjellman and Hudd, 1996.. In the flat terrain in the coastal areas of Finland, a variety of factors including natural phenomena Žisostatic land uplift. and human activities result in a need to dredge the middle and lower reaches as well as estuaries of many streams and rivers. This dredged material, which is rich in sulfide minerals and leachable heavy metals, is often deposited on the banks and, thus poses a potential threat to the aquatic environment. Hence, in addition to the extensive ditching of farmland, dredged masses most likely contribute to the acidification and metal loading of the surface waters in the coastal areas. The main aim of this study was to characterise, prior to the initiation of extensive dredging operations at many coastal sites, the geochemistry of lake sediments, which in the near future is likely to be dredged. The study consists of determination of concentrations, associations and spatial distribution of a variety of metals and non-metals
in such sediments, and an air-oxidation experiment in laboratory, by which the extent of acidity production and metal release on oxidation of the sediments can be assessed. We selected the southern part of the Larsmo lake Žmid-western Finland., which is an interesting area as it is strongly affected by isostatic land uplift Žapprox. 8 mmra. and was converted from a brackish-water bay to an artificial fresh-water lake in 1962.
2. Area description The data presented under this heading is mostly taken from Kalliolinna Ž1996.. The total drainage ¨ lakes is 4280 km2 area of the Larsmo and Oja ŽFig. 1.. The total area of the lakes at the fixed water level Ž N60 " 0.00 m. is 85 km2 . The area of ¨ lake 12 km2 . The Larsmo lake is 73 km2 and Oja total water volume is 200 million m3. The average ¨ water depth in Larsmo lake is 2.6 m and in Oja lake 1.6 m. Maximum water depths are less than 15 m. Four rivers discharge into Larsmo lake. Of these, the Esse river is largest Ždrainage area s 2054 km2 , median water flow 1989᎐1995 s 15.8 m3rs, account for 45% of the total inflow to the lake., followed by Purmo river Ž864 km2 , 6.9 m3rs, 25%., Kronoby river Ž788 km2 , 6.1 m3rs, 22%. and Kovjoki river Ž292 km2 , 2.4 m3rs, ; 10%.. The Esse and Purmo rivers have a joint mouth close to the lake ŽFig. 1.. The rivers are characterised by occasionally acidic waters rich in humic substances, Fe, N and P. The buffering capacity of the waters is low and occasionally almost absent. The Larsmo lake has had several mass kills of fish, which have been explained by ditching of farmlandrforests in the catchment and also to dredging of the river mouths. The fish mass kills have occurred mainly during spring and autumn high water flows. The local industries use a maximum of 5 m3rs of the fresh water from Larsmo lake and 1 m3rs ¨ lake. The remaining water in the lakes from Oja Žminus that volume which is evaporated. eventually discharges into the Botnian sea via locks at Hastgrundet and Gertruds ŽFig. 1.. The average ¨ outflow from Larsmo lake in 1989᎐1995 was 33
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
111
¨ lake. The sampling area Ža. and a dredging area Žb. are indicated. Fig. 1. Regional map over Larsmo-Oja
m3rs, of which 26 m3 Ž78%. went through Hastgrundet. The flow during this period ranged ¨ ¨ from 20.7 to 47.2 m3rs. The outflow points of Oja lake are at the Krakila ¨ ¨ and Palma locks ŽFig. 1.. These locks are, however, being used only at exceptionally high water levels. The outflow from ¨ ja lake is only 1% of that from Larsmo lake. O The annual amounts of material deposited in the lakes are as follows Ždata from 1989᎐1995.: 4100 t solids; 150 t nitrogen; and 14 t phosphorus. The annual material flow to the sea during this period was 6500 t solids, 940 t nitrogen and 48 t phosphorus. Of these amounts 30% comes from the Esse and Purmo rivers, 25% from the Kronoby river and 10% from the Kovjoki river. The material transported to the sea flows almost en-
tirely through Larsmo lake and only to some ¨ lake Ž1᎐2%.. extent via Oja
3. Methods 3.1. Sediment sampling Sampling of sediments at 39 sites was carried out during the summer of 1998 in the southernmost part of the Larsmo lake at the mouths of Esse, Purmo and Kovjoki rivers ŽFig. 1, area a.. The overall aim of the sampling was to collect the uppermost black sediment, which is present throughout most of the lake to depths of more than 2 m below the water-sediment interface.
112
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
These sediments consists of silt Žunpublished data of similar material from the same area.. The sample tool was a peat drill ŽRussian sampler., through which an undisturbed semicircular sample measuring 50 cm in length and 4.6 cm in radius is collected. A surface sample was taken from 0 to 50 cm and a deep sample from 50 to 100 cm. At three sampling points less than 1 m of the black mud was present and only the upper sample was taken, giving a total of 75 samples. The samples were immediately frozen in order to prevent oxidation. At the sampling points, the water depth varied between 0.2 and 2.5 m. All samples were black coloured with occasional occurrences of thin grey stripes, except two samples at the Kovjoki river mouth, which had a brown colour towards the surface. 3.2. Sediment analysis A portion of each sediment sample was oven dried below 60⬚C and comminuted in an agate mortar. A 2.0-g portion of the pulverised material was analysed for Al Ž100., As Ž2., B Ž3., Ba Ž1., Ca Ž100., Cd Ž0.01., Co Ž1., Cr Ž1., Cu Ž1., Fe Ž100., K Ž100., Mg Ž100., Mn Ž2., Mo Ž1., Na Ž100., Ni Ž1., P Ž10., Pb Ž2., Sr Ž1., V Ž1. and Zn Ž1. with ICP-AES after an aqua regia Ž95⬚C 2:2:2 HClrHNO3rH 2 O. digestion of the sample Žthe numbers in parenthesis are the detection limits in ppm.. This analysis was carried out in a commercial accredited ŽISO 9002. geochemical laboratory. In two samples Ž2r75. boron was below the detection limit. All other elements in all sediment samples were above the detection limits. Sulfur and carbon were determined in the same laboratory with Leco. The analyses of duplicates of 13 random samples Žnine for C and S. showed that for each element the data variance is significantly Ž99% confidence level. larger than the error variance. Aqua regia is a strong partial leach that mainly digests trioctahedral micas, clay minerals, salts and secondary precipitates ŽRaisanen, 1999.. ¨ ¨ 3.3. Sediment oxidation experiment and water analysis The oxidation experiment, was done under con-
trolled laboratory conditions and was designed to simulate sulfur oxidation and elemental leaching as they occur under field conditions. The experiment is a modification of a similar experiment ˚ ¨ and Bjorklund Ž1997.. For each done by Astrom ¨ sediment sample a portion of 3.0 g Ždry wt.. of reduced sediment was placed in a plastic tube and mixed with 9 ml of deionised water. The pH and conductivity were measured from the clear supernatants Žreducing conditions.. The samples were then kept at room temperature for 1 year and wetted with 3 ml of deionised water every month. After this period 9 ml of water was added and the tubes were shaken. After the sediment particles had settled, the pH and conductivity of the clear supernatants were measured Žoxidised state.. Control measurements were performed on the following day. In this experiment no leaching was allowed, since the added water Ž9 ml plus several 3-ml portions. was left to evaporate leaving the concentration of elements intact. After the measurements of the samples in the oxidised state, the supernatants were analysed for Al Ž1., As Ž1., Cd Ž0.02., Co Ž0.02., Cr Ž0.1., Cu Ž0.1., Li Ž0.5., Mn Ž0.05., Na Ž20., Ni Ž0.2., Pb Ž0.5., Rb Ž0.01., Sr Ž0.01., U Ž0.05. and Zn Ž0.5. with ICP-MS Žthe numbers in parentheses are the detection limits in ppb.. The analytical precision of the ICP-MS analysis was determined with in-house standard materials, reagent blanks and five replicates. The analytical precision was satisfactory and the ratio of data to error variance high for the determined elements. Chromium and As had 5r75 and 3r75 respectively, samples below detection. Some of the sediment samples were accidentally Žpartly. oxidised in the laboratory prior to the first measurement and their pH and conductivity values in reduced states was therefore not measured. These samples were not included in the statistical analyses. In the laboratory experiment, the waterrsoil ratio Žweight., with and without the monthly additions, were 13.3 and 3.0, respectively. In the field, the waterrsoil ratio for a period of 1 year Ž500 mm precipitation. is 1.78᎐2.23 and only 0.45 when the pore water has dried. Therefore, in the oxidation experiment there was excess water as com-
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
pared to the natural conditions. The temperature in the laboratory is approximately 20⬚C, while in the study area it is above 15⬚C on average for only 3 months per year ŽAtlas of Finland, 1987.. Hence, the results of the experiment most likely correspond to several years of oxidation under natural field conditions. The rate of oxidation in-field is of course affected also by numerous other factors such as the sulfur-bacterial activity, the thickness Žair penetration. of the sediments and climatic variables. 3.4. Statistical analyses The statistical analyses were performed with robust and non-parametric methods Žrange, median, Spearman correlation, Mann᎐Whitney Utest., which means that there is no requirement of normal data distributions. The data was examined for normal distribution with the Kolmogorov᎐ Smirnov test and with histograms. Approximately one half of the elements had a normal distribution Ž95% level., which motivates the use of nonparametric statistical methods. In most geochemical data analysis, non-parametric methods are preferred over those which require a normal data distribution ŽReimann and Filzmoser, 1999.. Growing symbol maps were prepared in order to graphically show the spatial variation in the data and to simplify the interpretation of possible geochemical distribution patterns.
4. Results and discussion 4.1. Element concentrations and associations The aqua regia extractable concentrations of elements in the Larsmo lake sediments are similar to those in 317 samples of similar sediments ˚ ¨ and Bjorklund, ŽAstrom 1997. distributed on the ¨ coastal plains of central western Finland ŽTable 1.. The relatively small, but significant, geochemical differences that exist between these two data sets are most likely related to different depositional environments, since while the sediments col-
113
lected from the coastal plains represent a heterogeneous group with an unknown mixture of lacustrine and marine sediments of varying age, the studied ŽLarsmo lake. sediments have been deposited within a geographically restricted area and are thus formed in a specific environment. In the Larsmo lake sediments, the concentrations of Na, B and S are significantly higher and Zn significantly lower in the deep Ž50᎐100 cm. than in the surface Ž0᎐50 cm. layer. The other elements do not occur in significantly different concentrations in the two layers ŽMann᎐Whitney U-test, 95% level.. The enrichment of Na and B in the deep layer is explained by a higher proportion of marinerbrackish Žcontains trapped NaCl and borate. to fresh water sediments in this than in the surface layer. This argument is in agreement with Eden ´ Ž1994., who found that the Na concentration in overbank sediments Ždeposited under fresh-water conditions. along 49 streams throughout Fennoscandia did not exceed 0.06%, a figure which is clearly lower than those in the deep layer of the Larsmo lake sediments Ž0.04᎐0.31%.. The ‘depletion’ of S in the surface layer is probably due to less reducing conditions here than deeper down in the sediment column. The maximum concentrations of the potentially toxic elements As, Ba, Cd, Co, Cr, Cu, Mo, Ni, Pb, V and Zn in the sediment samples were lower than the limit values for contaminated soils by factors of 3᎐100 ŽTable 1.. The only element that exceeded the limit value was S ŽTable 1.. The high S concentration is, however, considered normal since reducing conditions in the sediments have resulted in sulfide formation from sea-water sulfates and from sulfate derived from the extensively ditched S-rich fine-grained sediments in the drainage area ŽPalko et al., 1986.. High sulfur concentration is a typical feature also in many other reducing marine sedimenters elsewhere in the Botnian sea ŽMuller, 1999.. ¨ In the deep sediment layer, the elements can be divided into two main groups according to the interelement correlations and scattergram features: Ž1. Ca, K, Mg, Sr, B, Na, V, Cr, Ni, S and Cu; Ž2. Fe, Mn, Mo, Zn, Cd, Co, Ba, As, P, Pb and C. The highest coefficient for group 1 is
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
114
Table 1 Aqua regia extractable element concentrations in sediments in the southern Larsmo lake Žsurface and deep layer. and in ˚ ¨ 1998.. The last column shows the recommended corresponding sediments on the coastal plains ŽCP. of Finland Ž n s 317. ŽAstrom, and limit values for potentially toxic elements in soilsrsediments ŽAssmuth, 1997. Surface layer Ž0᎐50 cm.
Fe Ž%. C Ž%. Al Ž%. S Ž%. Mg Ž%. Ca Ž%. K Ž%. P Ž%. Na Ž%. Mn Žppm. Zn Žppm. Ba Žppm. V Žppm. Sr Žppm. Cr Žppm. Ni Žppm. Cu Žppm. Co Žppm. As Žppm. Pb Žppm. B Žppm. Mo Žppm. Cd Žppm.
Med
Max
Min
Med
Max
CP Ž100᎐350 cm. Med
Rec.rLim.
Min
Deep layer Ž50᎐100 cm.
1.46 1.18 0.56 0.16 0.25 0.27 0.12 0.08 0.02 220 35 26 19 15 13 8 7 6 4 5 bd 1 0.1
4.29 2.43 1.19 0.58 0.50 0.45 0.24 0.13 0.08 527 75 55 38 35 29 17 13 13 9 8 6 3 0.2
5.91 5.33 1.43 1.97 0.59 0.55 0.29 0.17 0.30 875 89 69 45 53 43 20 20 17 17 14 13 4 0.3
2.12 1.24 0.61 0.19 0.30 0.30 0.14 0.08 0.04 214 37 27 21 18 16 8 9 6 4 5 3 2 0.1
3.96 2.22 1.14 0.73 0.53 0.47 0.26 0.13 0.13 563 62 51 39 40 29 17 15 11 10 7 11 3 0.2
5.89 6.30 1.30 1.33 0.62 0.71 0.31 0.17 0.31 1176 93 66 46 66 55 19 22 16 20 16 15 4 0.3
3.80 1.37 2.02 0.54 1.04 0.50 0.60 0.06 0.10 448 90 87 50 36 48 31 27 13 ᎐ 12 ᎐ ᎐ ᎐
᎐ ᎐
recorded between Ca and Sr Ž rs s 0.93. and for group 2 between Co and Zn Ž rs s 0.95.. The highest coefficient recorded between one element from each group is that between V and Zn Ž rs s 0.40.. Aluminium has positive correlations with elements from both of the groups, but overall, higher with those of group 2. In the surface layer the correlations are generally similar except for the correlations involving elements which occur in considerably different concentrations ŽS, B and Na. in the two layers. For example, in the deep layer the Spearman correlation between Mn and S is y0.21 and in the upper layer 0.57, while for Fe and S the corresponding correlations are 0.0 and 0.55, respectively. The spatial geochemical patterns in the sediments in the southern Larsmo lake are affected by the general geochemical behaviour of the elements, the settling rate of different organic and
2 = 10y5 ᎐ ᎐ ᎐ ᎐ ᎐ ᎐ 90r700 600r600 50r500 ᎐ 80r500 40r300 32r400 50r200 13r60 38r300 ᎐ 5r200 0.5r10
inorganic colloidsrparticulates and by sediment resuspension and redeposition. Since the water depth in the study area is low Ž- 2.5 m. resuspension of the sediments may occur due to water currents, ice-cover movement and the frequent motor-boat traffic in summer. The spatial variations in the S concentrations in both sediment layers are shown in Fig. 2. Because the concentrations of the other elements were relatively low Žfar below limit values. and did not show any distinct spatial patterns, they are not shown. 4.2. Electric conducti¨ ity and pH The median pH redrox in the samples of the deep layer is 7.1r3.4 and the median conductivityredrox is 1055r3916 Srcm. In the surface layer the corresponding figures are 7.0r3.4 and 585r3035 Srcm, respectively. As these me-
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
115
Fig. 2. The distribution of sulfur in the sediments in the southern Larsmo lake Žthree dots are missing in the deep layer..
dian values and Fig. 3 show, the 1 year oxidation resulted in a substantial lowering of the pH values in all samples. This is caused by the oxidation of reduced sulfur by atmospheric oxygen, resulting in an extensive production of acids Žsulfuric acid and Fe. and thus a lowering of pH as the buffering capacity of the sediments is low Žlow contents of carbonate.. Higher sulfur concentrations generally produced lower pH ox , as is clearly shown by the correlations between these variables Ž rs s y0.50 and y0.56. ŽTable 2.. This relationship is further indicated by the fact that in 51% of the samples with the bottom 50% of the S concentrations the pH ox - 3.5, while for the samples with the top 50% of the S concentrations the corresponding figure was 71%. The concomitant strong increase in the median conductivity values Ž519% increase for the surface-layer samples and 372% increase for the deep-layer samples. shows that ions are formed and mobilised in large
amounts by the oxidation and acidification of the sediments ŽFig. 4.. The conductivity in the reduced state is controlled by the extent of trapped sea salts. This explains the higher conductivityred values of the deep-layer sediments, which in contrast to the surface sediments, to a large extent have been deposited in brackish Žand not fresh. water. 4.3. Mechanisms of metal release on oxidation The greatest relative leachability Žoxidation experiment. in the two layers are shown by Na Ž9.9᎐12.3%., Mn Ž9.6᎐9.7%., Cd Ž8.7᎐9.5%. and Co Ž4.3᎐5.2%. ŽTable 3.. Cobalt, Cd and Zn show an increase in the leached amounts for the surface layer ŽFig. 5.. This indicates that these metals are more weakly bound to the surface sediments, probably caused by antropogenic input in
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
116
Fig. 3. pH of the sediment samples in reduced and oxidised state. The loose dots represent the samples, which were accidentally exposed to air prior to the first measurements. The pH of these samples in the reduced state was therefore not measured.
recent decades. Another possible cause to the higher leachable fractions in the surface layer is the change from brackish to fresh water conditions 40 years ago. This decrease in salinity in the depositional environment during the last few decades explains the lower amounts of leachable Na in the surface layer.
The correlations between pH and the concentrations of Al, As, Cr, Cu, Li, Ni and U in the Žoxidised. leachates were inverse and high Žat least in one layer above 0.70, Table 4.. The critical pH values, which were obtained graphically from scattergrams and which indicate the pH point where a strong increase in the elemental
Table 2 Spearman correlation coefficients between the amounts of elements mobilised and the pH and conductivity developed by the oxidation of lake sediments Žfirst column., and the total S, total C and aqua regia extractable element concentrations in the sediments Sulfur
Carbon
Surface Na Al Mn Zn Sr Co Ni Cu Cd Cr Pb As Li Rb U pHox Condox U
U
0.49 0.56U 0.38U 0.42U 0.21 0.21 0.63U 0.70U 0.31 0.41U y0.18 0.36U 0.66U y0.44U 0.72U y0.50U 0.75U
Significant at the 95% level.
Deep y0.08 0.49U y0.25 y0.02 y0.05 0.20 0.61U 0.54U 0.01 0.38U 0.01 0.31 0.60U y0.48U 0.52U y0.56U 0.42U
Surface y0.24 y0.43U y0.22 y0.26 y0.40U y0.06 y0.58U y0.66U y0.35U y0.17 0.00 y0.45U y0.54U 0.19 y0.67U 0.43U y0.54U
Corresp. metal Deep 0.20 y0.29 0.10 y0.15 y0.05 0.04 y0.67U y0.66U y0.28 y0.33U y0.21 y0.44U y0.55U 0.54U y0.58U 0.42U y0.52U
Surface U
0.53 y0.20 0.91U 0.47U 0.46U 0.64U 0.52U 0.31 0.36U 0.06 0.02 0.31 ᎐ ᎐ ᎐ ᎐ ᎐
Deep 0.80U y0.23 0.87U 0.35U 0.63U 0.37U 0.30 0.29 0.18 y0.01 0.07 y0.12 ᎐ ᎐ ᎐ ᎐ ᎐
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
117
Fig. 4. Electric conductivity of the sediment samples in reduced and oxidised state. The loose dots represent the samples, which were accidentally exposed to air prior to the first measurements. The electric conductivity of these samples in the reduced state was therefore not measured.
concentrations begins were as follows: Ni 4.8; Cd 4.4; Co 4.0; Li 4.0; Al 3.9; As 3.9; Cu 3.8; U 3.8; and Cr 3.3 ŽFig. 5.. Zinc correlated relatively weakly with pH ŽTable 4. but the scattergram gives a critical pH value of approximately 4.8. Note, however, that in these results the pH spec-
trum is between 4.8 and 3.0, which means that the study of the elemental behaviour is limited to this pH range. Among the elements that correlated with the pH, three different patterns in the concentrationrpH scattergrams were observed. Ž1. The profiles for Al, ŽAs., Li, Cu and U are char-
Table 3 Easily mobilisable amounts Žppm. of metals and As Žas determined by the oxidation experiment. in two layers of lakermarine sediments. The last column Ž%. shows the relative amounts of easily mobilised elements as compared to the aqua regia extractable concentrations Surface layer Ž0᎐50 cm.
Na Al Mn Zn Sr Co Ni Cu Cd Cr Pb As Li Rb U
Deep layer Ž50᎐100 cm.
Min
Med
Max
%
Min
Med
Max
%
8.7 1.1 6.7 - 0.01 0.20 0.02 0.04 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01
83 71 55 2.7 1.4 0.65 0.47 0.06 0.02 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01
280 222 107 6.1 3.2 1.3 1.0 0.28 0.03 0.10 0.03 0.01 - 0.01 - 0.01 - 0.01
9.9 0.66 9.6 3.8 4.5 5.2 3.2 0.59 9.5 0.04 0.10 0.03 ᎐ ᎐ ᎐
25 0.94 2.7 0.14 0.30 0.04 0.08 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01 - 0.01
174 68 53 2.2 1.6 0.50 0.46 0.06 0.01 0.01 0.01 - 0.01 - 0.01 - 0.01 - 0.01
401 156 139 4.9 3.6 0.80 0.87 0.16 0.02 0.05 0.03 0.01 - 0.01 - 0.01 - 0.01
12.3 0.6 9.7 3.5 4.3 4.3 2.7 0.48 8.7 0.04 0.09 0.24 ᎐ ᎐ ᎐
118
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
Fig. 5. The relationship between pH Žand aqua regia extractable Mn concentrations . and selected element concentrations in the aquatic phase that had equilibrated with sediment samples oxidised for a period of 1 year in the laboratory.
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
119
Fig. 5. Ž Continued..
acterised by low concentrations at high pH and rapidly increasing concentrations at a given pH value, thus making it easy to obtain critical pH values ŽFig. 5.. Ž2. The concentrations of Cd, Co, Ni and Zn increase in a linear manner with decreasing pH values ŽFig. 5.. From these patterns it is more difficult to estimate a critical pH value since it probably lies at a higher pH than 5.0. Ž3. Chromium is the most immobile of the ‘pH sensitive elements’, showing a slight concentration increase only at very low pH values ŽFig. 5.. The elements whose mobility were least or not at all affected by the decreasing pH, in the observed pH range, were Na, Mn ŽFig. 5., Pb, Sr
and Rb. The concentration of Rb, in contrast to all other studied elements, increased in the solution phase as the pH of this became higher ŽTable 4.. Aluminium is derived from weathered silicates and is probably immobile as hydroxides and oxides at more neutral pH conditions. Al-hydroxysulfates in soils dissolve at pH - 4.2 ŽReimann and de Caritat, 1998.. Also Cr, Li and U have probably been bound as hydroxides and oxides. The elements As, Cd, Co, Ni and Zn probably derive from metal sulfides, which are stable at neutral and reducing conditions but which dissolves Žand oxidise. under aerobic conditions, to
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
120
Table 4 Spearman correlation coefficients between the amounts of elements mobilised and the pH and conductivity developed by the oxidation of sediments for 1 year pH
Na Al Mn Zn Sr Co Ni Cu Cd Cr Pb As Li Rb U U
Conductivity
Surface
Deep
Surface
Deep
0.23 y0.95U y0.24 y0.38U 0.01 y0.36U y0.65U y0.83U y0.31 y0.58U 0.38U y0.48U y0.8U 0.50U y0.79U
0.10 y0.93U 0.27 y0.42U 0.22 y0.54U y0.80U y0.89U y0.45U y0.76U y0.07 y0.80U y0.96U 0.73U y0.86U
0.49U 0.62U 0.60U 0.62U 0.50U 0.39U 0.76U 0.76U 0.58U 0.26U y0.28 0.48U 0.79U y0.57U 0.82U
0.68U 0.52U 0.12 0.36U 0.35U 0.20 0.65U 0.67U 0.33 0.34U y0.01 0.53U 0.64U y0.56U 0.67U
Significant at the 95% level.
form a metal Žor As. ion and sulfate. The results by Eden ´ Ž1994. indicate that a Na concentration above 0.06% in sediments is derived from trapped sea salts ŽNaCl.. The fact that the deep layer has higher easily leachable concentrations of Na than this means that the sediments have, to a greater extent, been deposited under brackish conditions. The leachable fractions of Mn Žhydroxides and oxides. probably dissolve at an earlier stage in the oxidation process, at pH values somewhere between 5 and 7. The soluble Mn fraction shows a strong correlation with the aqua regia extractable Mn concentration ŽTable 2, Fig. 5., which means that the proportion of easily soluble species Žrelative to aqua regia extractable Mn., such as Mn oxyhydroxides is constant throughout the sediments and controls the amounts of Mn leached. The pattern for Sr is somewhat similar to that of Mn, but the Spearman correlation for the leachable and aqua regia extractable Sr fraction is weaker than the corresponding one for Mn ŽTable 2.. Lead has probably been bound as sulfide in the sediments. Its leaching proportion is, however, relatively low ŽTable 3. which may be explained by sorption onto soilrsediment constituents after its release from the sulfides.
4.4. An assessment of en¨ ironmental impacts of dredging Since the isostatic land uplifting will continue for several thousand years in the coastal regions of central Finland and Sweden, dredging operations will certainly be frequent in these regions in the future, mainly in order to: Ž1. clear out sediments that accumulate at the river mouths, in which otherwise waterflow velocity will decrease with a concomitant risk of increased flooding; Ž2. deepen the sailing depths close to harbours; and Ž3. improve the recreational values along the shores of lakes and the coast. At present, a number of large dredging operations are planned in a variety of estuarinerlacustrine settings along the coasts. One of these is located east of the town of Jakobstad Žclose to the study area, Fig. 1, area b., where an area of 335 800 m2 of a shallow lake is to be dredged in the near future. In this area approximately 50 000 m3 of sediment will be dredged from the uppermost 1 m of the sediment column. The material to be dredged equals 23 300 t of sediment Ždry wt.. with a volume of 10 500 m3. Preliminary Žunpublished. analyses have shown that these sediments are geochemically similar to those of this study, except that they have even higher S concentrations Žmean s 1.0%, mean in this study s 0.71%. and thus probably consist an even larger environmental threat than the studied sediments. According to the median values Žppm. of the elements leached in the oxidation experiment ŽTable 3., these masses has the potential to release Žin the short term. considerable amounts Žkg. of Al Ž1710., Mn Ž1230., Zn Ž59., Sr Ž39., Co Ž13., Ni Ž12., Cu Ž2. and Cd᎐Cr᎐As᎐Pb Žless than 1 kg of each.. Because the dredged material will be deposited close to the site where it was collected, the mobilised elements could to a large extent end up in the lake, which is vulnerable to overloading due to its small water volume and low water circulation rates. While being somewhat speculative due to the large differences that exist between the conditions in field and laboratory, these figures highlight the fact that not only the potential acidity, but also the metals within the dredged masses pose a potential threat to the local aquatic environment. While these sediments
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
will be limed ŽCaCO 3 . and their pH thus kept relatively high at least in the short term, it is questionable whether such procedures will be enough to prevent acid and metal leakage in the long term because, Ž1. thorough liming of dense sediment masses is a difficult task, Ž2. if at some stage the acidity within the sediment becomes greater than the buffering capacity of the added lime, many ‘pH sensitive’ metals Že.g. Al, As, Cd, Co, Cu, Li, Ni, U. are likely to rapidly mobilise at decreasing pH ŽTable 4, Fig. 5. and Ž3. at the examined pH values ŽpH 4.8᎐3.0. some of the metals ŽMn and Sr. in the sediments are controlled by factors other than pH ŽFig. 5. and could thus become mobile even if the liming measures are successful. These figures and arguments thus suggest, that the potential acidity as well as metals within the sediments should be considered when measures are taken to minimise the environmental impact of the planned dredging operations.
5. Conclusions The concentrations of toxic elements in the Larsmo lake sediments are low and are as such no threat to the environment. However, the sediments contain large amounts of reduced sulfur, which is normal and natural, but which result in extensive acidity production and concomitant metal mobilisation on oxidation of the sediments. While under natural conditions, oxidation of the sediments occur at a low rate when the isostatic land uplifting exposes sediment layers to atmospheric oxygen, man made operations such as ditching and dredging quickly exposes large volumes of reduced sediments for the atmospheric oxidant. In the study area, the harmful effects of sediment oxidation is expected to be exceptionally high because of poorly buffered and biologically sensitive surface waters. In the highlighted dredging operation of 50 000 m3 of sediment, enough metals can be released to substantially damage the local aquatic environment. Consequently, in large dredging operations such as this one, care must be taken when selecting a site for dumping of the masses. Dumping close to
121
areas with high water circulation is recommended. If possible, the dredged material should be dumped as sub-surface fillings, which would prevent the oxidation of the reduced S. Dumping of dredged masses close to marine environments is preferred since marine water has far greater buffering capacity than fresh water. Neutralisation of dredged material is of great importance and is not to be neglected. While the oxidation experiment worked very well, it could still be improved to give more detailed results. For example, to allow for stepwise leaching where water is analysed continuously to get a more detailed pattern of acidity and metal mobilisation. The waterrsoil ratio could be optimised to be more close to that in field. To simulate the local climate may be unnecessary since the experiment would then be too time consuming. The sampling procedure was considered adequate and detailed enough for the purposes of the study. The differences in the Na concentrations in the layers highlighted the fact that Na Žand Cly. concentration can be used as a dating agent in this type of environment, if samples are collected at a much higher vertical frequency than was done in the present study. Acknowledgements The authors would like to thank for the financial support provided by the city of Jakobstad, UPM-Kymmene ŽJakobstad mills., Finska Vetenskaps-societenrSohlbergska delegationen and the Waldemar von Frenckel grant. The authors are also grateful to Soren Frojdo ¨ ¨ ¨ for help with the ICP-MS analyses and the figures, and to Peter ¨ sterholm for valuable advises during the project. O References ˚ ¨ M, Astrom ˚ ¨ J. Geochemistry of stream water in a Astrom catchment in Finland affected by sulfidic fine sediments. Applied Geochemistry 1997;12:593᎐605. ˚ strom A A. Impact of acid sulfate soils on stream ¨ M, Bjorklund ¨ water geochemistry in western Finland. J Geochem Explor 1995;55:163᎐170. ˚ strom A A. Geochemistry and acidity of ¨ M, Bjorklund ¨ sulfide-bearing postglacial sediments of western Finland. Environ Geochem Health 1997;19:155᎐164.
122
˚ ¨ r The Science of the Total En¨ ironment 284 (2002) 109᎐122 P. Peltola, M. Astrom
˚ strom A A. Hydrogeochemistry of a stream ¨ M, Bjorklund ¨ draining sulfide-bearing postglacial sediments in Finland. Water Air Soil Pollut 1996;89:233᎐246. ˚ strom A ¨ M. Mobility of Al, Co, Cr, Cu, Fe, Mn, Ni and V in sulfide-bearing sediments exposed to atmospheric O 2 : an experimental study. Environ Geol 1998;36:219᎐226. ˚ strom A ¨ M, Spiro B. Impact of isostatic uplift and ditching of sulfidic sediments on the hydrochemistry of major and trace elements and sulfur isotope ratios in streams, western Finland. Environ Sci Technol 2000;34:1182᎐1188. Atlas of Finland 1987. Climate folio 131. National Board of Survey and Geographical Society of Finland, 1987, pp. 32. Assmuth T. Analysis and proposals of guideline values for concentrations of hazardous substances in soil Žin Finnish.. Suomen Ymparisto ¨ ¨ Keskuksen Moniste 1997;92:56. Eden ´ P. Wide-spaced sampling of overbank sediment, till, humus and river water in Fennoscandia, PhD Thesis, Abo Academy University, 1994. Eden ´ P, Weppling K, Jokela S. Natural and land-use induced load of acidity, metals, humus and suspended matter in Lestijoki, a river in western Finland. Boreal Environ Res 1999;4:31᎐43. ¨ ¨ Kalliolinna M. Luodon-Ojanjarven vedenlaatu vuosina 1989᎐1995. The association of water protection in western Finland Žin Finnish., 1996, p. 52. Kjellman J, Hudd R. Changed length-at-age of burbot, Lola
lota, from an acidified estuary in the Gulf of Bothnia. Environ Biol Fish 1996;45:65᎐73. Muller A. Distribution of heavy metals in recent sediments in ¨ the Archipelago Sea of southwestern Finland. Boreal Environ Res 1999;4:319᎐330. ¨ ¨ Palko J., Rasanen M. and Alasaarela E. Luodon-Ojanjarven ¨¨ valuma-alueen maaperan ¨ ja vesiston ¨ happamuuskartoitus. Vesi ja ymparistohallituksen julkaisuja 1986, p. 11 Žin Fin¨ ¨ nish.. Palko J. Acid sulfate soils and their agricultural and environmental problems in Finland. Acta Univ. Oul. C 75, PhD Thesis, University of Oulu, 1994. Reimann C, de Caritat P. Chemical elements in the environment. Berlin Heidelberg New York: Springer-Verlag, 1998:398. Reimann C, Filzmoser P. Normal and lognormanl data distribution in geochemistry: death of a myth. Consequences for the statistical treatment of geochemical and environmental data. Environ Geol 1999;9Ž39.:1001᎐1014. Raisanen M-L. The podzolisation in Finland, processes and ¨ ¨ methods of analysis. Conference of Environmental Geology 23᎐24 March 1999, TurkurFinland Žin Finnish., 1999, pp. 39᎐43. Weppling K. Hydrochemical factors affecting the neutralization demand in acid sulfate waters. Vatten 1993;49:161᎐170.