DIOXIN AND POP-CONTAMINATED SITES • CHALLENGES • SERIES • RESEARCH ARTICLE

A decision framework for possible remediationof contaminated sediments in the River Kymijoki, Finland

Matti Verta & Hannu Kiviranta & Simo Salo &

Olli Malve & Markku Korhonen & Pia K. Verkasalo &

Päivi Ruokojärvi & Esko Rossi & Ari Hanski &Kare Päätalo & Terttu Vartiainen

Received: 24 June 2008 /Accepted: 3 October 2008 / Published online: 21 October 2008# Springer-Verlag 2008

AbstractBackground, aim, and scope The paper describes thespatial contamination of the River Kymijoki, South-EasternFinland, and the coastal region of the Gulf of Finland withPCDD/Fs and mercury. The findings of ecotoxicologial andhuman health studies are also reported, including environ-mental and human risk assessments. Sediments from theRiver Kymijoki, draining into the Gulf of Finland, havebeen heavily polluted by the pulp and paper industry and bychemical industries. A wood preservative, known as Ky-5,was manufactured in the upper reaches of the river between1940 and 1984 causing severe pollution of river sedimentswith polychlorinated dibenzo-p-dioxins (PCDD) and diben-

zofurans (PCDF). Moreover, the sediments have beenpolluted with mercury (Hg) from chlor-alkali productionand the use of Hg as a slimicide in pulp and papermanufacturing.Materials and methods An extensive sediment survey wasconducted as well as sediment transport modeling, toxicityscreening of sediment invertebrates, and a survey ofcontaminant bioaccumulation in invertebrates and fish.Studies on human exposure to PCDD/Fs and the possibleeffects on hypermineralization of teeth as well as anepidemiological study to reveal increased cancer risk werealso conducted. An assessment of the ecological and humanhealth risks with a null hypothesis (no remediation) wasundertaken.Results The sediment survey revealed severe contaminationof river and coastal sediments with PCDD/Fs and Hg. Thetotal volume of contaminated sediments was estimated toreach 5×106 m3 and hot spots with extremely highconcentrations (max 292,000 ng g−1 or 1,060 ng I-TEQg−1 d.w.) were located immediately downstream from thepollution source (approximately 90,000 m3). Sedimentcontamination was accompanied by changes in benthicassemblages, but direct effects were masked by manyfactors. The fish showed only slightly elevated PCDD/Flevels in muscle, but orders of magnitude higher in the livercompared with reference freshwater sites and the BalticSea. The concentrations in human fat did not reveal highhuman exposure in the Kymijoki area in general and waslower than in sea fishermen. The relative risk for totalcancer among farmers was marginally higher (RR=1.13)among those living close to the river, compared withfarmers living further away, and the possibility of increasedcancer risk cannot be ruled out. A conservative riskassessment revealed that the present probability of exceed-ing the WHO upper exposure limit of 4 pg WHO-TEQ kg−1

Environ Sci Pollut Res (2009) 16:95–105DOI 10.1007/s11356-008-0061-9

Series editors: Roland Weber, Mats Tysklind, Caroline Gaus

Responsible editor: Markus Hecker

Electronic supplementary material The online version of this article(doi:10.1007/s11356-008-0061-9) contains supplementary material,which is available to authorized users.

M. Verta (*) : S. Salo :O. Malve :M. KorhonenFinnish Environment Institute (SYKE),P.O. Box 140, 00251 Helsinki, Finlande-mail: matti.verta@ymparisto.fi

H. Kiviranta : P. K. Verkasalo : P. Ruokojärvi : T. VartiainenDepartment of Environmental Health,National Public Health Institute,P.O. Box 95, 70701 Kuopio, Finland

E. RossiEsko Rossi Oy,Kuokkasenmutka 4,40520 Jyväskylä, Finland

A. Hanski :K. PäätaloRamboll Finland Oy,Terveystie 2,15870 Hollola, Finland

d−1 for PCDD/Fs and DL-PCBs was 6%. The probability ofexceeding the WHO limit value of 0.23 μg kg−1 d−1 formethyl mercury was estimated to be notably higher at 62%.Based on these studies and the estimated risks connectedwith different remediation techniques a general remediationplan with cost benefit analysis was generated for severalsub-regions in the river. Dredging, on-site treatment, and aclose disposal of the most contaminated sediments(90,000 m3) was suggested as the first phase of theremediation. The decision regarding the start of remediationwill be made during autumn 2008.Conclusions The sediments in the River Kymijoki areheavily polluted with PCDD/Fs and mercury from earlierchlorophenol, chlor-alkali, and pulp and paper manufactur-ing. A continuous transport of contaminants is taking placeto the Gulf of Finland in the Baltic Sea. The highlyincreased PCDD/F and Hg levels in river sediments pose anecotoxicological risk to benthic fauna, to fish-eatingpredators and probably to human health. The risks posedby mercury exceed those from PCDD/Fs and need to beevaluated for (former) chlor-alkali sites and other mercuryreleasing industries as one basis for remediation decisionmaking.Recommendations and perspectives The studies form thebasis of a risk management strategy and a plan for possibleremediation of contaminated sediments currently underconsideration in the Southeast Finland Regional Environ-ment Centre. It is recommended that a detailed restorationplan for the most seriously contaminated areas should beundertaken. Based on current knowledge, the restoration ofthe whole river is not feasible, considering the current riskcaused by the contaminated sediment in the river and thecosts of an extensive restoration project. The experiencesgained in the present case should be utilized in theevaluation of PCDD/F- and mercury-contaminated sites inother countries. The case demonstrates that the historicreservoirs are of contemporary relevance and should beaddressed, e.g., in the national implementation plans of theStockholm Convention.

Keywords Chlor-alkali . Contaminated site management .

Mercury . PCDD . PCDF. Pulp and paper . Remediation .

Risk assessment . River Kymijoki . Sediment contamination

1 Background, aim, and scope

Recently, Weber et al. (2008a) provided an overview ofdioxin- (PCDD/F) and POP-contaminated sites and theircontemporary and future challenges. Inevitably, PCDD/Fand other POP-contaminated sites will remain a relevantenvironmental issue for future generations, because they posepossible environmental and human health risks. They need to

be secured and/or remediated and they need to be monitoredfor decades or even centuries. Weber et al. (2008a) estimatedthat the highest historical PCDDF and dioxin-like contam-ination globally has grown as a result of the production ofchlorinated chemicals and, in particular, including chlorinat-ed pesticides and PCBs. Furthermore, chlor-alkali productionand the pulp and paper industry have contributed to theglobal occurrence of contaminated sites (e.g., UNEP 2002;Munthe et al. 2007; Weber et al. 2008a). Due to thedimension, complexity and persistence of POPs, contami-nated mega sites and the immense costs of remediationmeasures, a comprehensive risk assessment involving mul-tidisciplinary research is necessary (Weber et al. 2008a).

This paper presents a case study of PCDD/F and mercury(Hg) contamination of an entire river system resulting fromchlorophenol and chlor-alkali production and from the use ofmercury as a slimicide in the pulp and paper industry.

A wood preservative, Ky-5, was manufactured in theupper reaches of River Kymijoki in the City of Kuusan-koski, South-Eastern Finland (Fig. 1) between 1940 and1984, from which an unknown amount of the product andits impurities ended up in the river and has been slowlytransported to the Gulf of Finland. The product consistedmainly of polychlorinated phenols (PCP), the main compo-nents being 2,3,4,6-tetrachlorophenol (83%), pentachloro-phenol (8%) and 2,4,6,-trichlorophenol (6%). PCDDs andPCDFs, especially higher chlorinated dibenzofurans oc-curred as impurities in the PCP product (Verta et al. 1999a,b). In a preliminary study, high levels of polychlorinateddibenzo-p-dioxins and dibenzofurans (PCDD/Fs) wereobserved in Kymijoki sediments with a maximum concen-tration of over 400,000 ng g−1 (Verta et al. 1999c). Hepta-and octachlorinated dibenzofurans, typical impurities of thetetrachlorphenol produced, represent 79% to 95% of thetotal PCDD/Fs in sediments the major congener being1,2,3,4,6,7,8-HpCDF.

A significant contribution from sources other than Ky-5 inthe River Kymijoki discharges seems unlikely due to thealmost complete absence of any congeners other than thosetypical in the Ky-5 product. Nonetheless, a contribution fromother sources that produce PCDF cannot be completely ruledout, especially from chlor-alkali production (Isosaari 2004).According to Rappe et al. (1991), the PCDF profile in chlor-alkali sludge is also dominated by certain 2,3,7,8, substitutedTCDF, PeCDF, HxCDF, and OCDF.

The sediments are also polluted with mercury (Hg) fromchlor-alkali production and the use of Hg as slimicide inpulp and paper manufacturing. More than 30 tons of Hgwas emitted between the 1940s and 1971 of whichapproximately two thirds originated from the pulp andpaper industry along the river stretch and one third fromchlor-alkali production at Kuusankoski (see Fig. 1). Highconcentrations of methyl mercury in fish have been

96 Environ Sci Pollut Res (2009) 16:95–105

detected and restrictions on fishing for human consumptionhave been in place since the late 1960s. During that period,the mean mercury concentrations in predatory fish (north-ern pike) in different river regions ranged from 1.9 to3.0 mg kg−1 fresh weight, (Kokko and Turunen 1988).Present concentrations in fish in the lower reach of the riverare still high, mean 0.82 mg kg−1 in pike, and less in non-predatory species (Rossi 2005).

Since 1996, multidisciplinary research has been carriedout in the River Kymijoki region as well as in the estuaryand in the Gulf of Finland to:

& Examine the spatial extent of the contamination& Elucidate and model the transport of the sediment and

pollutants& Predict future changes& Investigate ecotoxicological effects and assess risks on

aquatic biota& Assess human exposure and health impacts& Find out restoration needs and preconditions

This paper briefly describes the spatial contaminationin the river and the coastal region of the Gulf of Finland

and the findings of ecotoxicologial and human healthstudies including an environmental and human riskassessment. Previously, an assessment of the contributionof River Kymijoki derived contaminants to the sedimentsof the Gulf of Finland was conducted and compared withrecent data from other coastal and offshore sediments inthe Baltic Sea (Isosaari et al. 2002; Verta et al. 2007).These studies demonstrated the clear dominance of theRiver Kymijoki as a source of PCDD/Fs to the Gulf ofFinland compared to other sources. The studies form thebasis of a risk management strategy and a plan forpossible remediation of contaminated sediments currentlyunder consideration in the Southeast Finland RegionalEnvironment Centre (Fig. 2). To our knowledge this isone of the few comprehensive published risk assessmentsfor sediment remediation following pollution by thechlor-alkali/organochlorine industry (see Weber et al.2008a and references therein) and therefore address thisimportant issue here within the series on ‘Case Studies onDioxin and POP Contaminated Sites—Contemporary andFuture Relevance and Challenges’ in ESPR (Weber et al.2008a, b).

Fig. 1 The River Kymijoki study area with adjacent coastal region.The sampling sites (black spots) are defined only for the coastal sites.The statistics for PCDD/F and Hg concentrations in each sub region

(shaded) are shown in the boxes. The upper box represents thereference site upstream from the pollution source. All concentrationsare shown as dry weight

Environ Sci Pollut Res (2009) 16:95–105 97

2 Extent and burden of contaminated sediments

2.1 River system

The study area consists of a 130-km long river stretch withtwo main branches between Lake Pyhäjärvi and the Gulf ofFinland (see Fig. 1). The river water flow is regulated andflows through 11 power plants and six rapids. The totaldrop (water level between the first source and the Gulf ofFinland) is 50 m and the mean gradient is small (only

0.0006). The whole drainage area of the river system is37,200 km2 and 97% of the water running in the riverstretch under investigation from upstream the source area.Accordingly, only 3% is the direct discharge from theimmediate catchment near the source. The mean dischargeat the downstream end of the river is 330 m3 s−1.

The river bottom consists mainly of transport anderosion sites, which contain non-cohesive soil or solid clay,silt, and bedrock. The location of soft sediments, locatedmainly in the expansions of the river was estimated using

Fig. 2 A schematic diagram ofthe River Kymijoki risk assess-ment and remediation project

98 Environ Sci Pollut Res (2009) 16:95–105

echo and acoustic–seismic surveys. The material in sedi-mentation areas was usually gyttja, clay, and silt withvarying composition. Abundant wood debris and fibersoriginating from the pulp and paper industry was found inmost contaminated areas. Sediment cores were collectedfrom 1997–2003 through the soft sediment until the hardriver bottom or glacial clay layer was reached. Sedimentcores from 220 sites were sliced to 25-cm sub samples.Mercury concentration was analyzed from each sub sampleof 130 cores. PCDD/F concentration was analyzed fromsub samples of ten sediment cores and, additionally, from35 surface sediment samples from 15 locations in the riverand from ten coastal sites (see Fig. 1). Additional surveyswere conducted in the Gulf of Finland (Isosaari et al. 2002;Verta et al. 2007).

2.2 Spatial and temporal distribution of contaminants

Figure 1 shows the statistics of the concentration of PCDD/Fs(sum of all 2,3,7,8-substituted congeners and as I-TEQ) andmercury in sediments when the study area was classified intofive regions, the first being the reference upstream and the lastthe coastal region up to 70 km from the river estuary. The mostcontaminated area was immediately downstream from thepollution source with extremely high concentrations (max

292,000 ng g−1 or 1,060 ng I-TEQ g−1 d.w. at 0–25 cm deep;Salo et al. 2008). Generally, the maximum concentration wasdeeper in the profile but in some cases it was at the surfaceindicating the erosion of cleaner sediments deposited afterthe mid 1960s. The mean and maximum concentrations insediment profiles decreased downstream and were lowest inthe coastal sites farthest from the river estuary. In the coastalsites, the concentrations still clearly exceeded the back-ground concentration in offshore and reference areas for theGulf of Bothnia and the western Gulf of Finland withconcentrations generally below 0.5 ng g−1 d.w. (or 500 pgg−1; Verta et al. 2007).

The mercury concentration pattern followed that ofPCDD/Fs both regionally and mostly in temporal sedimentprofiles but with a more distinct decrease towards thesurface sediments compared with older and deeper sedi-ments. The maximum concentration 32 mg kg−1 wasmeasured close to a major emission source and where themaximum PCDD/F concentration was measured, butgenerally the sites in the near vicinity of the sourcesshowed the highest variability in concentration patterns.

The total burden of Ky-5 derived PCDD/Fs wasestimated as ca 5,960 kg (17.3 kg as I-TEQ) in the riverand 1,770 kg (12.4 kg as WHO-TEQ) in the Gulf ofFinland (Fig. 3). The total historical sum of approximately

Fig. 3 A schematic diagramshowing the burden of PCDD/Fs, total Hg and methyl mercury(MeHg) in different riverregions and annual transportbetween the pools and to theGulf of Finland in 2001. Thesize of the arrows is proportionalto the quantity

Environ Sci Pollut Res (2009) 16:95–105 99

29 kg (I-TEQ) is comparable with an approximately equalsum (20 kg I-TEQ) distributed with the Ky-5 product to thesawmill sites in Finland (Vartiainen personal communica-tion) and exceeds the annual atmospheric emissions of 17western European countries of approximately 13.7 kg/yearin 1985 and from 2.0 to 3.7 kg/year in 2005. Other PCDD/F-contaminated sites of this size (and larger) are presentedin Weber et al. (2008a,b). The total volume of contaminatedsediment (exceeding the level of 500 pg I-TEQ g−1 d.w.) inthe studied river course was 5×106 m3, the majority ofwhich was located in lake Tammijärvi (see Fig. 1). The 5-km long stretch immediately downstream from the Ky-5source contained approximately 23% of the total PCDD/Fburden in the river. Mercury was more evenly distributedwith only 5% situated in the vicinity of the major emissionsource (chlor-alkali plant) and the adjacent Ky-5 plant. Acomparison with the mercury emission estimate, above,shows that less than 10% of the mercury emitted histori-cally has been retained in the sediments and most mercuryhas been transported to the Baltic Sea (see Fig. 3).

3 Contaminant transport and future trends

Most of the PCDD/Fs eroding from the hot spot areaaccumulate in the lower course of the river. The annualPCDD/F flux to the Gulf of Finland has been estimated tovary from 13 kg to 24 kg/year (44–75 g/year I-TEQ; seeFig. 3). The annual Hg flux is from 26 to 32 kg/year. Basedon 1-year monthly monitoring an average of 30% mercuryis present in dissolved form. Earlier studies have shown thatthe River Kymijoki is the major source of PCDD/Fs in thesediments of the gulf, representing more than 90% of thetotal load of PCDD/Fs since the 1940s (Isosaari et al.2002). The estimate of the load to the gulf by Isosaari et al.2002 (1,770 kg or 12.4 kg WHO-TEQ, see Fig. 3.) may beunderestimated because certain lower chlorinated congenersare more water-soluble and have higher bioaccumulationand biomagnification potentials (e.g., 2,3,4,7,8-PeCDF and2,3,7,8,-TeCDF). Therefore, compartments other than sedi-ments may also be of importance in the total mass balanceof these congeners. A one-dimensional flow and transportmodel was used to simulate the transport of PCDD/Fcompounds and to assess the impact of a possiblerestoration dredging of the most contaminated sediments(Malve et al. 2003). Using the estimated trend in PCDD/Floading, downstream concentrations were calculated until2020. If contaminated sediments are removed by dredging,the temporary increase of PCDD/F concentrations indownstream water and surface sediments is estimated tostay within current guidelines of dredging in Finland. Anincrease of PCDD/Fs in the water would only be significantin the worst case scenario (10% resuspension of the

dredged sediment), but within annual variation in thepredicted case of 1–2% resuspension. Long-term predic-tions indicated only minor decrease in surface sedimentconcentrations with no remediation but a major decrease ifmost contaminated sediments close to the emission sourcewere removed. A pilot dredging operation was conducted in2001 to simulate the possible effects of dredging and tosupport model predictions. A test of four dredging methodsshowed that resuspension of less than 2% should beexpected based on current river characteristics and when acareful operation technique is used (Vesivalo et al. 2002).

A more detailed assessment of the effects of dredgingwas made using a two-dimensional sediment erosion andtransport model based on SMS (Surface-water ModelingSystem) for hydrodynamic modeling (U.S. Army Corps ofEngineers TABS-MD; Karvonen et al. 2005). This indicat-ed that with no remediation 25–50% of the dioxins and 20–40% of Hg deposited in the immediate sedimentary basinswould be transported downstream within the next 30 years.Removing and/or capping these sites (36,000 m3, 40% ofcontaminated sediments between the source and Keltti)would significantly reduce the PCDD/F transport down-stream. The highest transport percentages indicate circum-stances of increased high flows (probability of oneoccurrence within 100–500 years based on data from30 years of monitoring) because of climate change (see,e.g., Weber et al. 2008a). Accordingly, some 50% to 60%of the contaminants would not be flushed within the next30 years even during exceptionally high flows. However,the major fraction (70–80%) of Hg transport at the Kelttistation originates from additional sites and not from thePCDD/F-contaminated sites (30–40% from upstream, 40%from other, less contaminates sites) due to different locationof emission outlets (Salo et al. 2008). Consequently, theresidual contaminants for PCDD/Fs and Hg should beconsidered when evaluating management strategies.

4 Ecotoxicological effects

High toxicity effects of the sediment to exposed micro-organisms and high frequencies of mentum deformities inmidge (Chironomus spp.) larval populations were measuredin areas with high pollutant concentrations in sediments butoccasionally in the lower River Kymijoki as well. Thesediments did not clearly affect the survival of midge larvae,but growth and development tended to be slower in the mostcontaminated sediments (Verta et al. 1999a). Sedimentcontamination appeared to cause changes in benthic assemb-lages, but any direct effects on measured benthic parameterswere masked by many factors (Kiiski et al. 2005).

Highly elevated levels of polychlorinated diphenylethers (PCDE) that were also impurities of the Ky-5

100 Environ Sci Pollut Res (2009) 16:95–105

product, were also found in fish muscle below the pollutionsource but only slightly elevated PCDD/F levels (as I-TEQor WHO-TEQ) were present compared with fish from thesame species in Finnish freshwater and in the Baltic Sea.Ky-5 impurity congeners (e.g., 1,2,3,4,6,7,8-HpCDF) wereobserved in excess in fish from the river, but onlyoccasionally in the estuary (Korhonen et al. 2001). Becauseof the low TEF value of HpCDF, however, the increase inTEQ in river fish was only marginal. Baltic salmon andherring showed an order of magnitude higher level ofPCDD/Fs than River Kymijoki fish with a major congenerbeing 2,3,4,7,8-PeCDF, which is abundant in atmosphericsources and deposition but constitutes only a minor fraction(0.09% per cent) in the Ky-5 PCDD/F profile (Humppi andHeinola 1985; Vartiainen et al. 1995; Kiviranta et al. 2003).Concentrations in the fish’s liver below the Ky-5 sourcewere orders of magnitude higher than in muscle for most ofthe contaminants (Korhonen et al. 2001). The discrepancybetween liver and muscle was mostly due to the muchhigher fat content in the liver of these generally leanfreshwater fish species (Korhonen et al. 2001). The fishfrom the Baltic Sea also had a much higher fat contentcompared with the species inhabiting the River Kymijoki.

A comparison of the contaminant concentrations in thebenthic invertebrates with the concentrations in fish indicatedthat the contaminant bioavailability and bioaccumulation werenot directly proportional to concentrations of the contaminantsin sediments or at lower trophic level biota, and generallylower than expected. The PCDD/F congener composition infish was different to that in benthic invertebrates. For certain2,3,7,8-substituted PCDD/Fs biomagnification was observed(Lyytikäinen 2004).

Mercury concentrations in the fish varied spatially andamong species with the highest concentrations in largepredatory fish such as pike (Esox lucius), burbot (Lota lota)and large perch (Perca fluviatilis) with concentrationsgenerally above 0.5 mg kg−1 and occasionally more than1.0 mg kg−1 (Verta et al. 1999a). The highest concentrations

were measured in fish from the lower reach of the river andnot at the most polluted sites. Non predatory fish generallyhad concentration less than 0.5 mg kg−1. The ecotoxico-logical risk caused by mercury is mainly posed to fish-eating predators, especially otter (Lutra lutra; Rossi 2005).No analysis of mercury in mammals was available,however.

5 Human exposure, health, and epidemiological studies

5.1 Exposure studies

The exposure of the general population around the RiverKymijoki was studied analyzing human fat samples fromvolunteer patients in a hospital in Kotka, the largest cityclose to the river. Altogether, 390 human fat samples frompeople living close to the river the River Kymijoki wereanalyzed for PCDD/Fs. As a reference material, 420 humanfat samples were collected at the same time from 14different areas around Finland (Kiviranta et al. 2005). Theywere fat samples from volunteer appendicitis patientscollected for controls in the soft tissues sarcoma case–control study (Tuomisto et al. 2004). The concentrations inthe population from the area around the River Kymijokiwere at the same level as in the control samples whentaking into account that the Kymijoki population was, onaverage, 10 years older than the general population sampledfrom around Finland (Table 1).

Plasma fat concentrations of PCDD/Fs were measured infishermen from two locations in the Kymijoki area,representing inland lake fishermen and the Baltic-Sea-areafishermen. The concentrations of the sea-area fishermenwere about twice as high as in the lake area (210 and 110pg WHOPCDD/F-TEQ/g fat, respectively), indicating that theBaltic Sea is a more relevant source of PCDD/Fs whencompared to the River Kymijoki (Kiviranta et al. 2000;Kiviranta et al. 2002). A small additional exposure to PCDD/

Table 1 Median/range of age (years) and WHOPCDD/F-TEQ concentration (pg g−1fat) in different population groups in the River Kymijoki areaincluding reference data from the general population in Finland

Population group n Sampling Age WHOPCDD/F-TEQ Reference

Kymijoki general population 390 1999 52 (17–93) 34 (3–306) UnpublishedFinland general population 420 1997–1999 44 (13–81) 24 (4–153) Kiviranta et al. 2005Kymijoki fishermen Kiviranta et al. 2000

Kiviranta et al. 2002Inland 22 1997 60 (42–77) 110 (34–420)Baltic Sea 25 1997 59 (27–76) 210 (75–500)Finland fishermen 88 2004–5 (45–74) 94 Kiviranta et al. 2002Kymijoki breast milk 18 1997 27 (20–34) 12 (6–23) Hölttä et al. 2001Helsinki breast milk 14 1993–4 28 (19–36) 20 (8–34) Kiviranta et al. 1999bKuopio breast milk 28 1993–4 27 (18–39) 14 (6–26) Kiviranta et al. 1999b

Environ Sci Pollut Res (2009) 16:95–105 101

Fs originating fromKy-5 production could be obtained if largeamounts of River Kymijoki fish were used for nutrition(Kiviranta et al. 1999a). For comparison samples ofprofessional fishermen collected from other areas of theBaltic Sea around Finland are presented in Table 1. Theconcentrations among these fishermen were, on average,94 pg WHOPCDD/F-TEQ (Turunen et al. 2008). The timedifference between the sampling of the Baltic fisherman wasalmost 10 years which in part explains the lower concen-trations measured among Baltic Sea fishermen.

Mercury exposure studies were not conducted butprevious studies during the early 1990s revealed thatroughly 30% of humans with a high level of fishconsumption exceeded the WHO mercury concentrationguideline of 6 mg kg−1 Hg in hair (Alfthan 1995).

5.2 Health studies

To examine the prevailing levels of PCDD/Fs in humanmilk in the Kymijoki area, milk samples from 18 motherswere collected in 1997 (close to the Baltic Sea and in theinland river areas). Previous studies have shown theassociation between high dioxin exposure and demarcatedhypomineralization of teeth lesions (Alaluusua et al. 1999).Compared to the figures reported earlier in Finland, neitherthe prevalence of dental lesions nor the levels of PCDD/Fsin human milk from primiparae mothers (Kiviranta et al.1999b) were increased in residents of the studied areascompared to the general population (Hölttä et al. 2001).

5.3 Epidemiological studies

Relative risks (RR) for total cancer and 27 cancer subtypes(diagnosed in 1981–2000) were investigated in the Kymi-joki river area using regional health data, which holdsnationwide register data from the Population RegisterCentre, Statistics Finland and the Finnish Cancer Registry.Cancer incidence in all people (n=188,884) and in farmers(n=11,132) living closer than 20 km to the river was at thelevel expected based on national rates; so, no clear riskincrease was found for all residents. The relative risk fortotal cancer among farmers was marginally higher (RR=1.13) among those living <1.00 km from the river,compared with farmers living 5 to 20 km from the river(Verkasalo et al. 2004). A statistically significant increasewas observed for basal cell carcinoma of the skin (notincluded in the total cancers) in all residents living <5 kmfrom the river. Several other common cancers, includingcancers of the breast, uterine cervix, gallbladder, andnervous system, showed slightly elevated risk estimates at<5 km from the river. In this study, exposure assessmentwas based solely on the place of residence at one point intime. In other words, we calculated the distance between

the residence and the river shoreline but had no specificmeasure for PCDD/F exposure in participants. In generalmeasurement error tends to bias risk estimates towards thenull hypothesis suggesting that, most likely, the studywould have underestimated the risk of living near the river.On the other hand, the possibility of a chance effect,residual confounding by some SES-related lifestyle factor,or confounding by some unidentified factor cannot be ruledout even if the most representative reference population wasused (comparing farmers with farmers). Despite thelimitations of the exposure assessment, thus, the possibilitythat living close to the polluted river may have contributedto a small increase in cancer risk cannot be excluded,especially among farmers (Verkasalo et al. 2004).

6 Risk assessment

The aim of the ecological and health risks assessment wasto disclose the risks posed by the contaminated RiverKymijoki sediments based on the present situation (Rossi2005). No remediation scheme or action and their impactswere calculated or compared to the present risks. This was aconscious choice since a general remediation plan was notavailable and since a decision considering the necessity ofremediation actions was still pending. The risk assessmentcalculations included human and animal intake calculationsof PCDD/F and Hg. The risk assessment included dioxin-like PCBs (DL-PCB) as part of the background exposure.PCDE was also considered as a possible contaminant.Different intake route calculations were made using US-EPA procedures and the probability of exceeding the totaltolerable daily intake was estimated with a stochasticanalysis tool. The human risk assessment was conductedseparately for nine different river sub-regions with differentexposure levels. Fish eating, eating game birds, swimming,exposure to bank sediment through mouth and skin,inhalation and consuming cow milk (grazing on fields nearthe river bank) were included in the exposure assessment.

The average total daily intake of PCDD/Fs and DL-PCBs for a human living close to the most contaminatedsections of the River Kymijoki, swimming regularly in theriver, eating fish from the river exclusively, eating somegame bird and fish roe from the river area was estimated tobe 2.6 pg WHO-TEQ kg−1 d−1 from which DL-PCBsaccounted for 35% (Rossi 2005). This hypothetical life-style, truly an unrealistic one, was chosen to attain anestimate of the maximum risk, and to find out thesignificance of different exposure and intake routes. Themost significant single intake route was backgroundexposure, 0.98 pg WHO-TEQ kg−1 d−1 from which DL-PCBs accounted for 91%. It should be noted that the river isnot particularly polluted with PCBs although concentrations

102 Environ Sci Pollut Res (2009) 16:95–105

in fish are among the highest quartile in Finland. Theprobability of exceeding the WHO upper limit value of 4 pgWHO-TEQ kg−1 d−1 was estimated to be 6%. The highestintake route of PCDD/Fs from the river was fish consump-tion. The total dietary intake of the general Finnishpopulation is 1.5 pg WHO-TEQ kg−1 d−1 (including DL-PCBs; Kiviranta 2005).

The average total daily intake of Hg was estimated to be0.36 μg kg−1 d−1. The most significant single intake routewas eating fish, 0.27 μg kg−1 d−1. The probability ofexceeding the WHO limit value 0.23 μg kg−1 d−1 wasestimated to be 62% (Rossi 2005). New recommendationsfor eating fish in Finland released by the Finnish FoodSafety Authority in 2005 substantially reduce the risks ofexposure; for PCDD/Fs ca 20% and for Hg ca 56%. Ageneral consumer study in 2006 revealed that around 37%of consumers in Finland were aware of the recommenda-tion to limit consumption of large predatory fish becauseof high mercury concentrations. Local advice has beengiven to the public in the region, thus indicating betterthan average awareness of the recommendation. As awhole, the risks posed by Hg exceed those from PCDD/Fsand should not be ignored when making remediationdecisions.

7 General remediation plan as a decision-making tool

The main purpose of the general remediation plan (GRP)launched in 2007 was to review current informationnecessary for the remediation decision as briefly describedabove (Hanski and Päätalo 2007, see Fig. 2). Furthermore,different remediation options and techniques (mainlydredging and capping techniques) were examined. Risksconnected to different remediation techniques were com-pared, classified and ranked according to their anticipatedseriousness and probability (see also electronic supplemen-tal material—ESM). The possible risk reduction manage-ment options in each case were also suggested.

The possible risks included:

& Deficiency of information of geotechnical character-istics (affecting the success of the work)

& Effects on water quality& Erosion of suspended solids& Contaminant erosion and transport& Redistribution of contaminants in sediments& Mercury methylation during and after remediation& Redistribution of PCDD/Fs from particles to water& Behavior of other contaminants& Changes in river flow velocity& Changes in bottom ecosystem& Immediate effects on biota

& Effects on humans (through exposure to contaminants)& Effects on groundwater& Effects during sludge treatment and disposal

A cost benefit analysis (CBA) was performed for sevensub-regions in the river based on the simple cost efficiencyof removal or capping of contaminant (€/kg PCDD/F and €/kg Hg; more details in ESM). Dredging at site treatmentand a close disposal of the most contaminated sediments(90,000 m3) near the pollution source showed the best costefficiency, the anticipated costs being from 5.5 to 8.3million Euros (8.7 … 13.1 million US dollars). This wassuggested as the first phase of the remediation (Hanski andPäätalo 2007). The anticipated cost of the restoration for thewhole river was estimated to approximately 480 millionEuros (approx. 760 million US dollars). A preliminarystakeholder hearing of the plan was organized in 2008before the decision was made regarding continuation withthe detailed remediation plan of the most polluted area andenvironmental impact assessment procedure.

A pilot remediation project was launched in 2006–2007by encapsulating 2,300 m3 of contaminated sediments atone site representing an average flow velocity in the river.The increase of suspended solids downstream from the pilotand during the work varied from 3% to 60% compared withthe background value, the mean being 30%. Resuspensionfrom the sediment and from the capping material could notbe discerned (Kiirikki and Lindfors 2006). The study didnot support capping as a general remediation technique inthe river.

8 Conclusions

The sediments in the River Kymijoki are heavily pollutedwith PCDD/Fs and mercury from earlier chlorophenol-,chlor-alkali production, and pulp and paper manufacturing.The highly increased PCDD/F and Hg levels in riversediments pose an ecotoxicological risk to benthic fauna, tofish-eating predators and probably to human health. Therisks posed by mercury exceed those from PCDD/Fs. Acontinuous transport of contaminants to the Gulf of Finlandis taking place, and the River Kymijoki is the largest knowndioxin source to the Baltic Sea. Most of the PCDD/Fcompounds do not accumulate easily in fish and the dioxinlevels in fish in the Gulf of Finland are not elevatedcompared with other areas of the northern Baltic Sea. It isrecommended that a detailed restoration plan for the mostseriously contaminated areas should be undertaken. Basedon current knowledge, the restoration of the whole river isnot feasible, considering the current risk caused by thecontaminated sediment in the river and the costs of anextensive restoration project. The restoration is part of the

Environ Sci Pollut Res (2009) 16:95–105 103

Finnish national implementation of the Stockholm Conven-tion on POPs.

Suggested addition This Finnish national case shows thecontemporary relevance of a Dioxin/POPs and mercury-contaminated megasite from historic chlor-alkali/organo-chlorine industry. A decade long comprehensive researchprojects were necessary to sufficiently evaluate the rele-vance of the contamination and undertake a comprehensiverisk assessment to decide on remediation strategies.

It is recommended that countries having ratified theStockholm Convention address the historic Dioxin/POPsreservoirs/contaminated sites from chlor-alkali, organochlo-rine industries and other key sources (Weber et al. 2008a)within their national implementation plans in a systematicmanner and develop methodologies for this task asrequested by the Stockholm Convention.

Acknowledgement The Ministry of the Environment, Finland andUPM-Kymmene Corporation are acknowledged for their financialsupport and the South-Eastern Regional Environment Centre for theirsupport in field sampling.

References

Alaluusua S, Lukinmaa P-L, Torppa J, Tuomisto J, Vartiainen T(1999) Developing teeth as biomarker of dioxin exposure. Lancet353:206

Alfthan G (1995) Exposure to mercury from fish in high-fishconsumers in Finland (in Finnish). Publications of the NationalPublic Health Institute, Series B9/1995

Hanski A, Päätalo K (2007) Contaminated sediments in riverKymijoki—a general remediation plan (in Finnish). South-Eastern Finland Regional Environment Centre. Mimeographedreport 82116159, 1.6.2007, Ramboll Finland Oy

Hölttä P, Kiviranta H, Leppäniemi A, Vartiainen T, Lukinmaa P-L,Alaluusua S (2001) Developmental dental defects in childrenwho reside by a river polluted by dioxins and furans. ArchEnviron Health 56:522–528

Humppi T, Heinola K (1985) Synthesis and gas chromatographic–mass spectrometric determination of polychlorinated dibenzo-p-dioxins and related compounds in the technical chlorphenolformulation Ky-5. J Chromatogr 331:410–418

Isosaari P (2004) Polychlorinated dibenzo-p-dioxin and dibenzofurancontamination of sediments and photochemical decontaminationof soils. PhD thesis, University of Kuopio, Faculty of Natural andEnvironmental Sciences

Isosaari P, Kankaanpää H, Mattila J, Kiviranta H, Verta M, Salo S,Vartiainen T (2002) Spatial distribution and temporal accumulationof polychlorinated dibenzo-p-dioxins, dibenzofurans, and biphenylsin the Gulf of Finland. Environ Sci Technol 36:2560–2565

Karvonen T, Hepojoki A, Nurminen J, Verta M, Salo S, Malve O(2005) Using mathematical model to assess remediation optionsto dioxin- and mercury contaminated sediments in river Kymijoki(In Finnish). Mimeographed Report, Helsinki University ofTechnology and Finnish Environment Institute

Kiirikki M, Lindfors A (2006) Assessment of erosion of suspendedsolids to River Kymijoki from sediment construction at Koria

29.9.–20.10.1999. (In Finnish). Report of Luode Consulting Oy,4.12.2006

Kiiski A, Hämäläinen H, Salo S, Verta M, Kukkonen J (2005)Incidence of midge (Diptera, Chironomidae) larval deformities inthe assessment of biologically hazardous sediments. In: Meriläi-nen P, Sivula L, Oikari A. (eds), Seventh international conferenceof environmental sciences Jyväskylä, May 12–13, 2005. Sciencefor sustainability: Proceedings, Jyväskylä, Finnish Society forEnvironmental Sciences, University of Jyväskylä, p 225

Kiviranta H (2005) Exposure and human PCDD/F and PCB bodyburden in Finland. PhD Thesis, Publications of the NationalPublic Health Institute A 14/2005

Kiviranta H, Verta M, Korhonen H, Vartiainen T (1999a) Humanexposure to Ky-5 originating PCDD/Fs in Kymijoki area,Finland. Organohalogen Comp 44:169–172

Kiviranta H, Purkunen R, Vartiainen T (1999b) Levels and trends ofPCDD/Fs and PCBs in human milk in Finland. Chemosphere38:311–323

Kiviranta H, Vartiainen T, Verta M, Tuomisto JT, Tuomisto J (2000)High fish-specific dioxin concentrations in Finland. Lancet355:1883–1885

Kiviranta H, Vartiainen T, Tuomisto J (2002) Polychlorinated dibenzo-p-dioxins, dibenzofurans, and biphenyls in fishermen in Finland.Environ Health Persp 110:355–361

Kiviranta H, Vartiainen T, Parmanne R, Hallikainen A, Koistinen J(2003) PCDD/Fs and PCBs in Baltic herring during the 1990s.Chemosphere 50:1201–1216

Kiviranta H, Tuomisto JT, Tuomisto J, Tukiainen E, Vartiainen T(2005) Polychlorinated dibenzo-p-dioxins, dibenzofurans, andbiphenyls in the general population in Finland. Chemosphere60:854–869

Kokko H, Turunen T (1988) Mercury emissions to River Kymijokiand fish mercury concentrations up to 1986 (in Finnish).Vesitalous 3:30–38

Korhonen M, Verta M, Lehtoranta J, Kiviranta H, Vartiainen T (2001)Concentrations of polychlorinated dibenzo-p-dioxins and furansin fish downstream from a Ky-5 manufacturing. Chemosphere43:587–593

Lyytikäinen M (2004) Transport, bioavailability and effects of Ky-5and CCAwood preservative components in aquatic environment.PhD Thesis, Department of Biology, University of Joensuu,Finland

Malve O, Salo S, Verta M, Forsius J (2003) Modelling the transport ofPCDD/F compounds in a contaminated river and the possibleinfluence of restoration dredging on calculated fluxes. EnvironSci Technol 37:3413–3421

Munthe J, Bodaly RA, Branfireun BA, Driscoll CT, Gilmour CG,Harris R, Horvat M, Lucotte M, Malm O (2007) Recovery ofmercury-contaminated fisheries. Ambio 36:33–44

Rappe C, Kjeller L-O, Kulp SE, de Wit C, Hasselsten I, Palm O(1991) Levels, profile and pattern of PCDDs and PCDFs insamples related to the production and use of chlorine. Chemo-sphere 21:1629–1636

Rossi E (2005) An assessment of environmental and health riskscaused by the contaminated sediments of river Kymijoki (inFinnish). South-Eastern Finland Regional Environment Centre,Kouvola, Finland

Salo S, Verta M, Malve M, Korhonen M, Lehtoranta J, Kiviranta H,Isosaari P, Ruokojärvi P, Koistinen J, Vartiainen T (2008)Contamination of River Kymijoki sediments with polychlorinateddibenzo-p-dioxins, dibenzofurans and mercury and their transportto the Gulf of Finland in the Baltic Sea. Chemosphere Sept. 19,doi:10.1016/j.chemosphere.2008.07.085>

Tuomisto JT, Pekkanen J, Kiviranta H, Tukiainen E, Vartiainen T,Tuomisto J (2004) Soft-tissue sarcoma and dioxin: a case-controlstudy. Int J Cancer 108:893–900

104 Environ Sci Pollut Res (2009) 16:95–105

Turunen A, Verkasalo PK, Kiviranta H, Pukkala E, Jula A, Männistö S,Räsänen R, Marniemi J, Vartiainen T (2008) Mortality in a cohortwith high fish consumption. Int J Epidemiol 37:1008–1017

UNEP (2002) Global mercury assessment. Report by UNEP Chem-icals, Geneva, Switzerland, December 2002, p 129, http://www.chem.unep.ch/mercury/

Vartiainen T, Lampi P, Tolonen K, Tuomisto J (1995) Polychlorodi-benzo-p-dioxin and poly-chlorodibenzofuran concentrations inlake sediments and fish after a ground water pollution withchloro-phenols. Chemosphere 30:1439–1451

Verta M, Ahtiainen J, Hämäläinen H, Jussila O, Järvinen H, KivirantaH, Korhonen M, Kukkonen J, Lehtoranta J, Lyytikäinen M,Malve O, Mikkelson P, Moisio V, Niemi A, Paasivirta J, Palm H,Rantalainen A-L, Salo S, Vartiainen T, Vuori K-M (1999a)Organochlorine compounds and heavy metals in the sediment ofRiver Kymijoki; occurrence, transport, impacts and health risks(In Finnish). Publications of the Finnish Environment Institute,No 334

Verta M, Korhonen M, Lehtoranta J, Salo S, Vartiainen T, Kiviranta H,Kukkonen J, Hämäläinen H (1999b) Ecotoxicological and healtheffects caused by PCPs, PCDEs, PCDDs, and PCDFs inKymijoki sediments, South-Eastern Finland. DIOXIN 99. 19thInternational Symposium on halogenated environmental organicpollutants and POPs. September 12–17, 1999. Venice, Italy.Organohalogen Comp 43:239–242

Verta M, Lehtoranta J, Salo S, Korhonen M, Kiviranta H (1999c) Highconcentrations of PCDDs, and PCDFs in River Kymijoki

sediments, South-Eastern Finland, caused by wood preservativeKy-5. DIOXIN 99. 19th International Symposium on halogenat-ed environmental organic pollutants and POPs. September 12–17, 1999. Organohalogen Comp 43:261–264

Verta M, Salo S, Korhonen M, Assmuth T, Kiviranta H, Koistinen J,Ruokojärvi P, Isosaari P, Bergqvist P-A, Tysklind M, Cato I,Vikelsøe J, Larsen MM (2007) Dioxin concentrations in sedi-ments of the Baltic Sea—a survey of existing data. Chemosphere67:1762–1775

Verkasalo PK, Kokki E, Pukkala E, Vartiainen T, Kiviranta H,Penttinen A, Pekkanen J (2004) Cancer risk near a polluted riverin Finland. Environ Health Persp 112:1026–1031

Vesivalo J, Salo S, Mykkänen E, Aunola T, Hietapakka U, Malve O,Verta M (2002) Test dredging of contaminated sediments inKymijoki river in 2001 (in Finnish). Regional EnvironmentalPublications; 286—Kouvola: Southeast Finland Regional Envi-ronment Centre, Finland, 2002

Weber R, Gaus C, Tysklind M, Johnston P, Forter M, Hollert H,Heinisch E, Holoubek I, Lloyd-Smith M, Masunaga S, MocarelliP, Santillo D, Seike N, Symons R, Torres JPM, Verta M,Varbelow G, Vijgen J, Watson A, Costner P, Woelz P, Wycisk P,Zenneg M (2008a) Dioxin- and POP-contaminated sites—contemporary and future relevance and challenges. Environ SciPollut Res 15:363–393

Weber R, Tysklind M, Gaus C (2008b) Dioxin—contemporary andfuture challenges of historical legacies. Environ Sci Pollut Res15:96–100

Environ Sci Pollut Res (2009) 16:95–105 105