Document y8E5mZxg2JMZR2eVm8zMnDXD

ELSEVIER The Science of the Total Environment 290 (2002) 181-198 the Science ofthe total Environment All InlermtOoml Ju'inial fir .Sitr.illfk Kntanh into Iht EnvIroBBfnl and lit Rclaliamhlp "Oh Man www.elsevier.com/locate/scitotenv Towards a global historical emission inventory for selected PCB congeners -- a mass balance approach 1. Global production and consumption Knut Breivik3 *, Andy Sweetmanb, Jozef M. Pacyna3, Kevin C. Jonesb "Norwegian Institute for Air Research, P.O. Box 100, N-2007 Kjelier, Norway bEnvironmental Science Department, Institute of Environmental and Natural Sciences, Lancaster University, Lancaster, LAI 4YQ, UK Received 15 May 2001; accepted 15 October 2001 Abstract Information on the historical global production and consumption of polychlorinated biphenyls (PCBs) is urgently needed for estimating PCB fluxes to the environment and for interpreting global contamination patterns by these pollutants. This study presents the methodology, principal uncertainties and selected results from an inventory, aiming to quantify the global production and consumption of total PCBs as well as 22 PCB congeners. The available data on the historical production of PCBs and the chemical composition of various technical mixtures have been compiled from the literature. For some producers with less detailed information, the production of individual PCB constituents has been estimated to derive a global estimate for individual homologues and selected congeners. Information on imports, exports and consumption, as well as restrictions on production and imports, has further been compiled for individual countries. These data, along with assumptions on the trade between countries and regions, have been utilised to derive an estimate of the global historical consumption pattern. Although there are substantial uncertainties involved in these estimates, important aspects governing the large scale temporal and spatial patterns are most likely captured in these estimates. In particular, the information on imports and exports for the principal users of PCBs around the time of peak production is considered to be fairly reliable. The estimates account for a reported historical global production of --1.3 million t PCBs, more than 70% of which are tri-, tetra- and pentachlorinated biphenyls. The results further suggest that almost 97% of the global historical use of PCBs have occurred in the Northern Hemisphere. 2002 Elsevier Science B.V. All rights reserved. Keywords: POPs; PCBs; Homologues; Congeners; Sources; Production; Consumption; Global; Historical 1. Introduction Environmental contamination by polychlorinat- *Corresponding author. Tel.: +47-63-89-8000; fax: + 47 63-89-8050. E-mail address: knut.breivik@nilu.no (K. Breivik). ed biphenyls (PCBs) was recognised more than 30 years ago when Soren Jensen detected PCBs in pike from Sweden (Jensen, 1966). Since then, numerous studies have detected PCBs in various compartments of the environment (e.g. Edwards, 1971; Kalmaz and Kalmaz, 1979; Waid, 1986) and the occurrence of PCBs in remote areas, such 0048-9697/02/$ - see front matter 2002 Elsevier Science B.V. All rights reserved. P1I: S0048-9697(01)01075-0 LEXOLDMON003383 WATER_PCB-00000013 J 82 K. Breivik et al. / The Science of the Total Environment 290 (2002) 18J~-]98 as the Arctic (e.g. AMAP, 1998) is evidence for the long-range atmospheric transport of these con taminants (Oehme andMano, 1984; Oehme, 1991; Hamer et al., 1998). Today, PCBs are considered an environmental problem of global proportions. Even though the production of these contami nants has stopped, PCBs continue to be detected in environmental samples from around the world (e.g. Iwata et al., 1994; AMAP, 1998). Many studies have sought to understand historical PCB contamination trends through the analysis of dated sediment (e.g. Christensen and Lo, 1986; Sanders et al., 1992; Bruckmeier et al., 1997) and peat cores (Rapaport and Eisenreich, 1988). Several of these studies have indicated that the trends in environmental concentrations have followed the trends in production and use of PCBs. Historical data on the global production and usage of PCBs are, thus, urgently needed for the interpretation of historical, present and future contamination levels around the world (Cummins, 1988; Tanabe, 1988; Voldncr and Li, 1995; Wania and Mackay, 1996; Vallack et al., 1998). Furthermore, quantitative knowledge of the global historical consumption is a prerequisite for estimating atmospheric emissions and eventually establishing source-receptor rela tionships for intentionally produced PCBs on a global scale. Due to differences in lifetime and fate of individual PCBs in the environment, esti mates of the historical production of PCBs have to be done on a congener specific basis. The overall aim of this study was to present a quanti tative estimate of the historical consumption of selected PCB congeners. Specifically, we set out to: 1. estimate the historical global production of selected PCB congeners (temporal pattern); 2. estimate the historical global pattern of con sumption (spatial pattern); and 3. provide input for a global PCB emission model, presented in an accompanying paper (Breivik et al., 2002). 2. Methods The general molecular formula for the PCBs is C]2H]0--f,Cl,,, where n could be any number from 1 to 10. There are, thus, 10 different PCB homo- logues, dependent on the number of chlorines and 209 different PCB congeners, dependent on the position of the chlorines on the molecule. Not all congeners have been identified in commercial products or technical mixtures. The numbering system proposed by Ballschmiter and Zell (1980) has been adapted by the International Union of Pure and Applied Chemists (1UPAC), and is fre quently used to refer to various congeners. In this system, individual PCB congeners arc assigned a number, ranging from PCB-1 (2-CB) to PCB-209 (2,2',3,3',4,4',5,5',6,6'-CB). This numbering sys tem is also used here, although with minor revi sions (Hillery et al. 1997). In this work, 22 individual PCB congeners were studied. These are the same congeners as selected in the EU Global- SOC project (ENV4-CT97-0638), or more specif ically PCB-5 (2,3-DiCB), PCB-8 (2,4'-DiCB), PCB-18 (2,2',5-TriCB), PCB-28 (2,4,4'-TriCB), PCB-31 (2,4',5-TriCB), PCB-52 (2,2',5,5'-TetCB), PCB-70 (2,3',4',5-TetCB), PCB-90 (2,2',3,4',5- PenCB), PCB-101 (2,2',4,5,5'-PenCB), PCB-105 (2,3,3',4,4'-PenCB), PCB-110 (2,3,3',4',6-PenCB), PCB-118 (2,3',4,4',5-PenCB), PCB-123 (2',3,4,4',5-PenCB, PCB-132 (2,2',3,3',4,6- HexCB), PCB-138 (2,2',3,4,4',5'-HexCB), PCB- 149 (2,2',3,4',5',6-HexCB), PCB-153 (2,2',4,4',5,5'-HexCB), PCB-158 (2,3,3',4,4',6- HexCB), PCB-160 (2,3,3',4,5,6-HexCB), PCB- 180 (2,2',3,4,4',5,5'-HepCB), PCB-194 (2,2',3,3 ',4,4 ',5,5 '-OctaCB ) and PCB-199 (2,2',3,3',4,5,5',6'-OctaCB). We proceeded by first collecting from the liter ature data on the production of total PCBs as well as of various technical PCB mixtures. Secondly, these data were combined with data on the com position of these technical mixtures to estimate the production of individual homologues and conge ners. To fill gaps in the data, assumptions had to be made sometimes concerning the homologue and congener composition (Section 3.1). The global consumption pattern was assessed by compiling information on imports, exports and consumption of PCBs for individual countries and years. Reli able information is available only for countries with a historically high consumption of PCBs. For LEXOLDMON003384 WATER_PCB-00000014 K. Breivik et at. / The Science of the Total Environment 290 (2002) 181-198 183 Table 1 Total PCB production in t as reported in the literature Producer Monsanto Geneva Ind. Kanegafuchi Mitsubishi Bayer AG Prodelec S.A. Cros Monsanto Caffaro Chemko Orgsteklo Orgsintez Xi'an Total Country USA USA Japan Japan West Germany France Spain UK. Italy Czechoslovakia USSR (Russia) USSR (Russia) China Start 1930 1971 1954 1969 1930 1930 1955 1954 1958 1959 1939 1972 1960 1930 Stop 1977 1973 1972 1972 1983 1984 1984 1977 1983 1984 1990 1993 1979 1993 Amount 641 246 454 56 326 2461 159 062 134 654 29 012 66 542 31 092 21 482 141 800 32 000 8000 1 324 131 Reference de Voogt and Brinkman (1989) de Voogt and Brinkman (1989) Tatsukawa (1976) Tatsukawa (1976) de Voogt and Brinkman (1989) de Voogt and Brinkman (1989) de Voogt and Brinkman (1989) de Voogt and Brinkman (1989) de Voogt and Brinkman (1989) Schlosserova (1994) AMAP (2000) AMAP (2000) Jiang et al. (1997) other countries, assumptions had to be made on the trade between various countries and regions, using the Gross Domestic Product as a surrogate parameter (Section 3.2). 3. Results and discussion 3.1. Global production 3.1.1. PCB production process The production of PCBs involves the chlorina tion of biphenyl in the presence of a catalyst. Depending on the reaction conditions, the degree of chlorination varies between 21% and 68% chlorine on a weight-by-weight basis (e.g. Ahlborg et al., 1992). The homologue profile for most technical formulations shows a normal distribution around the mean chlorine content (e.g. Takasuga et al., 1996, see also Table 2). This implies that these mixtures generally contain only a certain `range' of PCB homologues and congeners as indicated by the chlorine content. However, some other technical formulations, such as Aroclor 1232, do not show this typical distribution, suggesting that they consist of more than one technical mix ture (Frame, 1997). 3.1.2. Total global PCB production A review of the literature data was undertaken to obtain the most reliable production figures for the major producers of PCB in various countries. A previous study had estimated the cumulative global production to be on the order of 1.5 million t (de Voogt and Brinkman, 1989). Most of the information presented in Table 1 is adapted from this compilation of data. The figures shown in Table 1 add up to a reported total global production of 1.324 million t between 1930 and 1993. Most likely the true cumulative global production has been higher, as there were factories in Poland (Falandysz et al., 1992), Eastern Germany (de Voogt and Brinkman, 1989) and Austria (Fiedler, 1997) that produced PCBs in unknown amounts. Although the data presented here might be lower than the real figure, it seems likely that most of the global historical production is accounted for in these estimates. For most producers within the OECD countries, data on the total amounts pro duced are generally reported for 5-year periods from 1955 to 1984. In addition, annual production data are available for the same countries from 1973 to 1980 (de Voogt and Brinkman, 1989). For producers outside the OECD, there is only limited information on the annual production from various plants. Production data reported for a period in excess of one year (e.g. a 5-year period) were uniformly distributed over that period (temporally flat distributed, see also Fig. 3a). These data show that Monsanto (USA) has been responsible for LEXOLDMON003385 WATEFLPCB-0000001 184 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 Fig. 1. Overview of the relationship between the annual pro duction of total PCBs and the production of individual homologues and congeners from various producers. Details of the approach are discussed in Section 3.1.3. The homologue composition (by wt.%) of var ious technical mixtures is given in Table 2, along with the estimated maximum and minimum default homologue composition. The annual productionweighted congener default compositions were esti mated in a similar manner based on data of the technical mixtures produced by Monsanto (USA) and Bayer (West Germany) (data not shown). As the availability of data for the different producers varied immensely, an individual approach was necessary to maximise the utilisation of available information and to minimise the use of uncertain assumptions. For the sake of transpar ency, details of the approach adopted for each individual producers are given in the following paragraphs. Although this might appear unneces sary, we hope that such detailed documentation will eventually facilitate future improvements to these estimates. almost 50% of the known reported historical pro duction of PCBs. Bayer (West Germany), Prodelec (France) and Orgsteklo (Russia) have each con tributed individually with more than 10% of the historical production. 3.1.3. Estimated production of individual homo logues and congeners The overall methodology for estimating the production of individual PCB homologues and congeners is depicted in Fig. 1. Whenever possible, information on the production of individual tech nical mixtures (e.g. Aroclor 1242), and their chem ical composition was used to determine the homologue and congener production over time. However, as this information was not available for many producers, the annual production of individ ual homologues and congeners could only be estimated in this manner for Monsanto (USA) and Bayer (West Germany) (see below and Fig. 1). For other cases, we estimated a set of annual production-weighted default compositions (i.e. a fraction between 0 and 1). These default compo sitions or fractions take into account the temporal shift in homologue and congener production that occurred during the period of peak production (i.e. they were varying from 1955 to 1983). 3.1.3.1. USA. The data compiled by de Voogt and Brinkman (1989) account for a total production of 641 699 t of PCBs in the USA. Monsanto was by far the most important producer, while the only other PCB-producing company in the USA (Gene va Industries) has a reported production of only 454 t. The two last numbers in the names of the technical Aroclor mixtures by Monsanto refer to the weight percentage of chlorine (Table 2). For example, Aroclor 1260 should contain approxi mately 60 wt.% chlorine. Two notable exceptions to this numbering system are Aroclor 1016, which is a technical mixture derived by distillation of Aroclor 1242, and Arcoclor 1232 which is a blend of approximately equal proportions of Aroclors 1221 and 1242 (Frame, 1997). The congeneric composition of Aroclors 1016, 1221, 1232, 1242, 1254, 1260 and 1262 were based on data compiled by Frame (1997), who reported the results of a collaborative study in which the composition of six Aroclor mixtures (1016, 1221, 1242, 1254, 1260 and 1262) was determined using 18 gas chromatographic systems. Nine systems employed ECD detection, and the remaining half-used MS- LEXOLDMON003386 WATER_PCB-00000016 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 185 Table 2 Content of the 10 homologues of PCBs in various technical mixtures as used in the calculations (in wt.%) Technical mixture Aroclor 1221 Aroclor 1232 Aroclor 1016 Aroclor 1242 Aroclor 1248 Aroclor 1254 Aroclor 1260 Aroclor 1262 Clophen ABO Clophen A40 Clophen A50 Clophen A60 Sovol TCB Delor 123 Delor 103 Default compositions Min Max Mono- 43.8 26.5 0.7 0.3 0.2 - 9.0 1.0 0 1.0 Di- 27.9 23.9 17.1 14.7 2.5 0.5 0.1 0.2 19.6 0.2 - - 14.0 63.0 10.0 6.6 19.6 Tri- 4.4 27.0 53.6 42.1 22.8 0.7 0.3 1.2 48.1 17.3 0.2 1.0 49.0 26.0 60.0 19.5 48.1 Tetra- 2.6 18.7 27.7 33.9 51.7 18.3 0.9 1.1 25.0 50.5 17.6 0.7 23.0 32.0 2.0 26.0 22.9 28.4 Penta- 0.5 3.5 0.8 8.1 20.5 55.6 9.9 3.9 6.2 25.8 51.2 16.2 53.0 4.0 3.0 6.2 22.5 Hexa- 0.2 0.3 0.1 0.8 2.0 22.0 43.5 28.1 1.2 5.0 26.8 49.1 22.0 1.0 - 1.2 17.0 Technical mixture Hepta- Octa- Nona- Deca- Notes Aroclor 1221 Aroclor 1232 Aroclor 1016 Aroclor 1242 Aroclor 1248 Aroclor 1254 Aroclor 1260 Aroclor 1262 Clophen A30 Clophen A40 Clophen A50 Clophen A60 Sovol TCB Delor 123 Delor 103 Default compositions Min Max - 0.1 - 0.1 0.3 2.5 36.1 45.3 1.2 3.6 27.8 1.0 - 0 9.3 -- -- --0.1 0.4 ~ 8.3 0.9 18.5 1.7 --0.6 5.8 0.5 -- ---- 00 2.1 0.2 -A -A - A,B - A,B -A - A,B - A,B -A -B -B -B -B -C -C -D -D 0E 0E The original data from the literature were scaled to yield 100%, except Aroclor 1221 (see text). [A] Frame (1997); [B] Schulz et al. (1989); [C] Ivanov and Sandell (1992); [D] de Voogt and Brinkman (1989); [E] Annual production-weighted compositions (or ratios) are based on the estimated production from Bayer AG (West Germany) and Monsanto (USA). Only the minimum and maximum compositions are shown. SIM or full-scan MS ion-trap measurements (see Frame, 1997 for details). Average weight percent age of individual congeners in each of these Aroclors was given for both ECD and MS systems. Frame (1997) also reported data on the composi tion of Aroclor 1232 and 1248, analysed by one of the participating laboratories in his study (referred to as JWC). In addition, data for Aroclor 1016, 1242, 1254 and 1260 from a previous study (Schulz et al. 1989) were included. We used a weighted average of these data to calculate the congeneric composition of the Aro- LEXOLDMON003387 WATER_PCB-00000017 186 K. Breivik et at / The Science of the Total Environment 290 (2002) 181-198 clor production. These compositions, reported in Table 2, are thus based on 19 different measure ments for Aroclor 1016, 1242, 1254 and 1260, 18 different determinations for 1221 and 1262, and a single determination for 1232 and 1248. These data were also used to estimate the homologue composition of each individual mixture and scaled to yield a total of 100%. The composition of the lighter chlorinated Aroclor 1221 was not scaled this way, because it presumably contains signifi cant amounts of biphenyl due to incomplete chlorination. The amounts of various Aroclors (1016:1242:1248:1254:1260 and `Other Aroclors') sold in the USA from 1957 to 1975 are available from de Voogt and Brinkman (1989). These data were used to estimate and scale the production of individual Aroclors, assuming that the production of various Aroclors was equal to the annual frac tion of sold amounts. For the years prior to 1957 and after 1975, we used the estimated fractions for 1957 and 1975, respectively. The sold amounts of the `Other Aroclors' (1221 + 1232+1262+1268) was generally below 3% during the investigated time-period (de Voogt and Brinkman, 1989). In the absence of information on the chemical com position of Aroclor 1268, we assumed that `Other Aroclors' was a mixture of 1221, 1232 and 1262 (1:1:2). As the `Other Aroclors' include the heav ily chlorinated mixture 1268 in unknown relative amounts, it is likely that we are underestimating the produced amounts of some of the more chlo rinated PCB homologues and congeners. Monsanto also produced a technical mixture called Aroclor 1270 (de Voogt and Brinkman, 1989). Due to the lack of data on both production and chemical composition, it could not be included in this estimate. For the production of PCBs by Geneva Indus tries we relied on the default homologue and congener composition (see Fig. 1). 3.1.3.2. West Germany. PCBs were produced in West Germany by Bayer AG as Clophens (A30, A40, A50, A60). The data presented by de Voogt and Brinkman (1989) account for a historical production of 159,062 t PCB. According to this reference, the approximate wt.% of chlorine in various trade mixtures were: A30 (40-42%), A40 (48%), A50 (52-54%) and A60 (60%). Fiedler (1997) presented production data by degree of chlorination (39, 42.5, 47, 48.5, 54, 55 and 60% Cl (w/w)) for the period from 1974 to 1983. As the amounts of individual Clophens produced were not available to us, we assigned the data presented by Fiedler (1997) to the corresponding Clophenmixture, based on the degree of chlorination in order to estimate fractions of the technical mixtures produced annually. For data prior to 1974, we assumed that the various mixtures were produced in the same relative quantities as in 1974. The homologue and congener production could then be estimated based on the homologue and congener content of A30, A40, A50 and A60 reported by Schulz et al. (1989). 3.1.3.3. Japan. Approximately 96% of the total Japanese PCB production was by Kanegafuchi Chemical Co. Ltd (Tatsukawa, 1976), which pro duced a series of PCB mixtures called Kanechlors (KC). According to Tatsukawa (1976), the chlo rine content of KC-300, KC-400, KC-500 and KC600 corresponds to the Aroclors 1242, 1248, 1254 and 1260, but the homologue composition is dif ferent. Although the chemical composition of Kanechlors has been at least partly determined (Kannan et al. 1992; Takasuga et al., 1996), this information could not be used because the individ ual amounts of Kanechlors produced arc unknown to us. Therefore, we used annual production data for various homologues reported by Tatsukawa (1976) for the years 1961 to 1971 as tri-CBs and lower, tetra-CBs, penta-CBs and hexa-CBs and higher. For the two clustered homologue groups (tri-CB and lower, hexa-CB and higher), we assumed that the internal homologue production was determined by the percentage of possible congeners. For example, the group of tri-CB and lower includes 39 possible congeners, while there are three possible congeners within the group of mono-CB. Thus, 7.7% of the clustered tri-CBs and lower was assumed to be mono-CB, etc. The production of individual congeners was estimated using the congeneric default composition (Fig. 1). These percentages were multiplied by the reported or estimated annual production of individ LEXOLDMON003388 WATER_PCB-00000018 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 187 ual homologues as described above. This proce dure clearly introduces high uncertainties to the estimated annual production of individual conge ners, but it seems reasonable to assume that this approach avoids emphasis on congeners that are rarely formed during the production process. For the other producer in Japan (Mitsubishi), we applied default homologue and congener compo sitions to total PCB production figures reported by Tatsukawa (1976). 3.1.3.4. Czechoslovakia. Schlosserova (1994) reported that 14 140 t of PCB was produced in Czechoslovakia as Delor 103. In addition, 4381 t of Delor 106 (similar to Aroclor 1260) and 2961 t of other PCB mixtures were produced from 1959 to 1984. de Voogt and Brinkman (1989) reported that approximately 6000 t were produced annually before 1968, suggesting higher production figures than those reported by Schlosserova (1994). How ever, we have chosen to use the more detailed and recent information provided by Schlosserova (1994), assuming a uniform annual production of 826 t total PCBs throughout the period. To estimate the homologue production, we used the homologue composition of Delor 103 as report ed by de Voogt and Brinkman (1989) and assumed that Delor 106 has the same composition as Aro clor 1260 (Table 2). For the remaining 4381 t, we assumed that its composition is that of Delor 123, a very light PCB formulation described by de Voogt and Brinkman (1989). We further used the congeneric default composition for estimating the production of individual congeners. 3.1.3.5. United Kingdom. In the UK, PCBs were manufactured under the trade name Pyroclor (de Voogt and Brinkman, 1989). As we did not have reliable information on the relative production volume, or chemical composition of the Pyroclors, we applied the default homologue and congener compositions. The UK factory was owned by Monsanto Industrial Chemicals Co., which also was the major producer of PCBs in USA. Similar ities with the production process at Monsanto in the US are thus likely. 3.1.3.6. Soviet Union. Estimates of the historical production of PCBs in the former Soviet Union are available from Ivanov and Sandell (1992) and more recently from AMAP (2000). According to this latter source, PCBs were produced under three different brand names (Sovol, Sovtol and TCB) at two different factories in the vicinity of Moscow (Orgsteklo and Orgsintez). Sovol is reported to have a chemical composi tion fairly close to that of Aroclor 1254 (Ivanov and Sandell, 1992; Takasuga et al. 1996). How ever, another study suggests that it only resembles Aroclor 1254 to a limited extent (Kannan et al. 1992). According to AMAP (2000), 43 000 and 9500 t of Sovol were produced at Orgsteklo (1939-1990) and Orgsintez (1972-1993), respectively. Sovtol (Soviet oil) has been characterised as a mixture of Sovol and trichlorobenzene (Ivanov and Sandell, 1992; AMAP, 2000). In particular, Sovtol-10 is a mixture of 90% Sovol and 10% trichlorobenzene (AMAP, 2000). Although other Sovtol-mixturcs are known (Ivanov and Sandell, 1992), we assumed that all Sovtol contained 90% Sovol. According to AMAP (2000), a total of 32 000 t of Sovtol were produced at Orgsteklo (1939--1987) and another 25 000 t at Orgsintez (1972-1990). This results in an estimated total production of Sovol-based PCBs of 103 800 t. This figure corresponds well with the 100 000 t previously estimated by Ivanov and Sandell (1992). Ivanov and Sandell (1992) also refer to another formulation produced in the former USSR, named `Trichlorodiphenyl' (TCDP), stating that this was only a product name and that its chemical com position was fairly close to that of Aroclor 1242. We assume this to be the same formulation as TCB; Trichlorobiphenyl in AMAP (2000). This assumption seems reasonable in light of the fact that both sources list the same production period and use as a dielectric fluid. According to AMAP (2000), 70 000 t of TCB were produced at Orgs teklo during 1968-1990. This value is also used here, although it is considerably higher than the previous uncertain estimate of 25 000 t given by Ivanov and Sandell (1992). The homologue composition of Sovol and TCB was taken from Ivanov and Sandell (1992) and is given in Table 2. For the congener production, we LEXOLDMON003389 WATER_PCB-00000019 188 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 assumed Sovol to be equal to Aroclor 1254 and TCB to be equal to Aroclor 1242. 3.1.3.7. China. PCBs were neither produced nor used in large amounts in China (Xu et al., 2000). According to Jiang et al. (1997), only 8000 t of PCB were produced in China during the 1960s and 1970s. The two principal technical formula tions are said to be similar to Aroclor 1242 and Aroclor 1254 (Jiang et al., 1997). In order to estimate the homologue and congener production, we assumed that these two technical mixtures were produced in equal amounts and had the chemical composition of the above mentioned Aroclors. 3.1.3.8. France, Spain and Italy. The data from France, Spain and Italy include total production figures and detailed annual records for the years 1973 1984. Otherwise, only the factories and trade names, as well as a few data on the homo logue and congener content of some of the tech nical mixtures were found in the literature (de Voogt and Brinkman, 1989; Kannan et al. 1992). As no data on annually produced amounts of various mixtures or constituents were available to us, the estimated production of individual homologues and congeners had to rely on the default homologue and congener compositions. 3.1.4. Uncertainties in the global production pattern The methods applied to estimate production rates of individual homologues and congeners rest on a number of critical assumptions. The use of the default homologue and congener composition (e.g. for France, Spain and Italy) introduces uncer tainties that are difficult to quantify. Obviously, the use of these default values implies that there were similarities in the production process between various producers, and implicitly, that there is a different propensity among the congeners to be formed during the production process. For some countries (e.g. China, USSR, Czechoslovakia), similarities have been noted in the chemical com position of technical mixtures from different pro ducers. It thus seems reasonable to assume that there have indeed been similarities in the produc tion process between different producers, resulting in similarities in the composition of the mixtures, at least at the homologue level (see, e.g. Takasuga et al. 1996). We are aware that there are substantial variations in the propensity of congeners to be formed during the manufacturing process (Frame, 1997). As the selected approach based on default compositions was designed to capture this varia tion, we only estimated production-weighted com positions based on technical mixtures that had been completely characterised (Schulz et al. 1989; Frame, 1997). In any case, the available data make it difficult, if not impossible to quantify all ofthe uncertainties in a quantitative and meaningful way. For example, it is likely that the composition of the same technical mixtures varied, at least to some extent, from batch to batch (de Voogt and Brinkman, 1989; WHO, 1993). Secondly, the relative produc tion rates of various mixtures were in many cases extrapolated based on reported data available for a few years only (e.g. Bayer AG). However, the impact of the use of homologue and congener compositions can be addressed in a simplified manner. The same applies for the vari ability between Aroclor compositions determined with either ECD or MS systems (see Frame, 1997). Therefore, we tried to quantify and depict these two sources of uncertainty. Table 3 summarises how uncertainty was addressed for the production by various producers. Details are explained in the following. At the homologue level (Table 3), the analytical variability (i.e. the difference between the homo logue content of the Aroclors determined by ECD and MS) is estimated for the Aroclor production at Monsanto (USA). For most producers, the maximum and minimum homologue compositions are used to estimate some of the anticipated vari ations at the homologue level. Kanegafuchi (Japan) was treated as a special case, accepting the estimates for tetra- and penta-CBs as reported, and applying max/min compositions as deter mined specifically for this particular producer. The uncertainties at the homologue level for Bayer AG, Chemko, Orgsteklo, Orgsintez and Xi'an were not addressed because we anticipate that the assump tions made are of less quantitative importance than the variability in these compositions. LEXOLDMON003390 WATER_PCB-00000020 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 189 Table 3 Overview of the selected approach to estimate the uncertainties associated with the use of homologue and congener compositions as well as the variation in the characterised mixtures of Aroclors Homologue level Congener level Analytical variability Monsanto (USA) Monsanto (USA) Orgsteklo Orgsintez Xi'an Max/min compositions Geneva Industries Kanegafuchi Mitsubishi Prodelec S.A. Cros Caffaro Monsanto (UK) Geneva Industries Kanegafuchi Mitsubishi Prodelec S.A. Cros Caffaro Monsanto (UK) Chemko Not addressed Bayer AG Chemko Orgsteklo Orgsintez Xi'an Bayer AG At the congener level, we used the analytical variability for Monsanto (USA) as well as for other producers for which the congeneric compo sition of the production was deduced from the Aroclors (Orgsteklo, Orgsintez and Xi'an). For most other companies, we used the maximum and minimum congener compositions multiplied with the corresponding max/min homologue composi tions. The uncertainties at the congener level some how represents a `worst case' (max/max and min/ min compositions). For Bayer AG, the uncertainty has not been addressed, due to lack of useful quantitative information for this purpose. Fig. 2 presents results for the estimated total global production of individual homologues (A) and 22 selected congeners (B). The results indicate that of the 1324-kt total PCBs accounted for 566 kt (42.7%) can be attributed to the 22 selected congeners. Fig. 2 also indicates that tri-CBs have been the most important PCB homologue produced historically. Whereas there are theoretically 24 trichlorinated congeners, ranging from PCB-16 to PCB-39, a closer inspection reveals that approxi mately 50% of this homologue group can be attributed to only three congeners: PCB-18, PCB28 and PCB-31. Similarly, the relative contribution of the other selected congeners to their respective homologue groups are 49% (Di-CBs), 24% (TetraCBs), 50% (Penta-CBs), 59% (Hexa-CBs), 22% (Hepta-CBs) and 43% (Octa-CBs). Table 4 pres ents the estimated global production rates of indi vidual homologues for various time-periods (in percent). As can be seen from Table 4, the global homologue production pattern has changed over time. Notably, there was a reduction in the relative contribution of PCBs with seven or more chlorines over the last few decades that PCBs were pro duced. This is easily explained as the production of heavier technical mixtures decreased in the USA and West Germany in the last years of production (USA from 1970, West Germany from 1974) as a result of an increased environmental awareness towards the heavier (more persistent) homologues and congeners. For example, the estimated relative importance of Hexa-CB produced by Monsanto (USA) decreased by approximately 50% from the 1960s to the 1970s. Furthermore, the production of Hepta-CBs and Octa-CBs by Monsanto (USA) essentially ceased after 1973. This has profound influence on the applied homologue and congener compositions and their time dependence. Thus, Table 4 reflects the assumption that a similar decrease in the production of the more chlorinated LEXOLDMON003391 WATER_PCB-00000021 190 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 Fig. 2. Estimated global production of individual PCB homologues (A) and congeners (B) in thousands of tons. congeners occurred worldwide around the same time. Unfortunately, this cannot be verified with the available information. Fig. 3 presents results for the estimated temporal pattern in the global production of total PCBs (A), PCB-28 (B), PCB-52 (C), PCB-101 (D), PCB118 (E), PCB-138 (F), PCB-153 (G), PCB-180 (H) from 1930 until 1993. The compiled data suggest a peak annual production of 75.5 kt total PCBs for the year 1970. Presumably, worldwide production of PCBs ended in 1993 when the production of Sovol ceased in Russia (AMAP, 2000). The results indicate further that the time trend of the production of individual congeners resembles that of the total PCBs. There are, how ever, notable exceptions when the uncertainties are taken into consideration. For the heavier congeners (Fig. 3F-H), the uncertainty is relatively high for the years after 1970. Again, this is a reflection of the use of the (default) homologue and congener compositions and their time dependence. As a result, the uncertainty for the later years of pro duction increases. The negligible uncertainty depicted for individual congeners prior to 1955 are not due to the availability of more reliable data for this period. Rather, these historical data reflect that there were only a few producers worldwide and the production was dominated by Monsanto (USA) and Bayer (West Germany). The possibility to present any meaningful quantitative estimate of the uncertainties during that time period is consid ered limited. LEXOLDMON003392 WATER_PCB-00000022 K. Breivik et al. / The Science of the Total Environment 290 (2002) 18J-J98 191 Table 4 Estimated global production rates of individual homologues (in percent) and sum of all homologues (in kt) for various time-periods Period 1930-1934 1935-1939 1940-1944 1945-1949 1950-1954 1955-1959 1960-1964 1965-1969 1970-1974 1975-1979 1980-1984 1985-1989 1990-1993 Mono- 0.3 0.3 0.3 0.3 0.3 0.5 0.7 0.8 0.8 0.2 <0.1 0 0 Di- 8.6 8.2 7.2 7.2 7.2 7.5 8.3 10.3 11.1 10.3 13.7 7.3 6.3 Tri- 25.0 24.1 21.1 21.1 21.1 21.5 23.4 28.7 32.0 29.8 37.2 26.0 22.5 Tetra- 24.5 24.4 24.2 24.2 24.2 23.5 24.2 26.9 24.9 24.9 26.0 27.7 27.0 Penta- 16.0 17.4 22.0 22.0 22.1 21.4 20.5 18.2 17.5 21.3 16.0 27.5 31.0 Hexa- 14.3 14.6 15 5 15.5 15.5 15.3 13.8 9.4 9.4 10.7 6.2 11.1 12.6 Period 1930-1934 1935-1939 1940-1944 1945-1949 1950-1954 1955-1959 1960-1964 1965-1969 1970-1974 1975-1979 1980-1984 1985-1989 1990-1993 Hepta- 9.1 8.8 7.8 7.8 7.7 8.2 7.3 4,4 3.4 2.3 07 0.5 0.6 Octa- 2.1 2.0 1.7 1.7 1.7 1.9 1.7 1.0 0.8 0.4 0.1 0 0 Nona- 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.1 <0.1 <0.1 <0 1 0 0 Deca- <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0 0 Sum homologues (kt) 36.4 37.8 43.5 43.5 43.8 97.7 168.4 285.6 274.7 169.5 86.6 29.2 6.8 3.2. Global consumption The global consumption of PCBs has been estimated based on information about import, export and national consumption, as well as restric tions on imports of PCBs in various countries and regions. In the absence of detailed information, which would facilitate an analysis for individual technical mixtures, we generally assumed that the import and consumption of individual congeners are deter mined by the production within representative countries or regions. This seems to be the only approach feasible with respect to the available data. Perhaps the most serious limitation in this methodology is that PCBs were exported as tech nical mixtures, rather than as individual congeners, as it is well-known that PCBs were sold according to the physical properties of the technical mixtures (WHO, 1993). Hence, some countries most likely had a relatively high import of one type of tech nical mixture (or product containing one type of a mixture). However, by treating groups of countries as closed markets, potential important regional variations in the spatial and temporal patterns of homologue and congener consumption can be addressed. An overview of the method to address the global consumption pattern is given in Fig. 4, while details of the approach are discussed below. 3.2.1. OECD countries -- consumption and export Detailed data on imports and exports from OECD-countries, along with data on exports to OECD and non-OECD countries, are available for the period 1973-1980 from de Voogt and Brink man (1989). Tatsukawa (1976) reported annual data on import and export for the entire period of LEXOLDMON003393 WATER_PCB-00000023 192 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 1930 1940 1950 1960 1970 1980 1990 2000 1930 1940 1950 1960 1970 1980 1990 2000 Fig. 3. Estimated temporal trend in the global production of total PCBs (A); PCB-28 (B); PCB-52 (C); PCB-101 (D); PCB-118 (E); PCB-138 (F); PCB-153 (G); and PCB-180 (FI) in thousands of tons. the PCB production in Japan. A key assumption for the national consumption within the OECD countries was that the OECD was a closed market. Whereas export from OECD was allowed to occur, import of PCBs into the OECD from non-OECDcountries was considered negligible. This should be a valid assumption in terms of quantitative importance, even though exceptions are likely. Furthermore, we assumed that prior to 1946, no export from producing countries occurred. For example, PCBs have been imported and used in Norway since approximately 1950, according to the national environmental protection authorities (SFT, 1996). The annual national consumption of PCB-congeners in OECD-countries in the 1970s was esti mated in the following way. de Voogt and Brinkman (1989) reported the annual exports from producing countries within the OECD to both OECD and non-OECD countries (see Fig. 4). The total annual export amount was thus divided into two sums, one for distribution to OECD countries and one for non-members. Next, we utilised infor mation on annual imports of PCBs for specific countries within OECD for the same time-period (de Voogt and Brinkman, 1989). For all producing OECD-countries, there were data on imported amounts for most reference years of the given LEXOLDMON003394 WATER_PCB-00000024 K Breivik et at. / the Science of the Total Environment 290 (2002) 1 S' 1 -19H 193 PRODUCERS Production within OECD-countries CONSUMERS ; Export to OECD j : countries ! Export to non- j OECD countries [ Producing OECD countries Non-producing OECD countries "Other countries" Czechoslovakia Eastern Europe (50%) Czechoslovakia (50%) USSR China {production ~ consumption) Fig. 4. Overview of the spatial distribution of global PCB con sumption. Details of the approach are discussed in the text. time-period, except for a few years for which we estimated the import with linear interpolation. Imports to all non-producing OECD-countries were based on annual reported imports whenever possible. The `excess export amount' to OECD countries that could not be accounted for, was distributed to the remaining OECD countries based on gross domestic products (reference year 1983) (United Nations, 1994). An easy approach to evaluate this assumption would be to compare our estimates with independently derived estimates on national PCB consumption from e.g. national envi ronmental agencies (see discussion later). Finally, for producing OECD-countries, we utilised either national consumption data or added the sum of import plus the amounts produced but not attrib uted to export, to estimate the national annual consumption. For the years from 1946, up to the time when unifying statistical data became available, we had to make some simple assumptions, except for Japan where detailed data are available (Tatsukawa, 1976). We assumed that producing OECDcountries were exporting the same relative amounts of the total production as reported for the early 1970s. We used the gross domestic products of the OECD countries to distribute the estimated total export to OECD-countries. Canada was treated differently as its total cumulative import had been previously estimated to be approximately 40 kt (de Voogt and Brinkman, 1989). In our calculation we assumed a steady linear increase in the import from 1946 until 1972, withdrawing the amount imported to Canada in the 1970s. Concerning the export to non-OECD countries (see Fig. 4), we assumed a linear increase in the fraction of export from 1946, until export-figures from OECD-producing countries to non-OECD countries became available for the early 1970s. This approach leads to uncertainties in the esti mated national consumption data, and it is difficult to validate the results. However, the estimated data are comparable to some other independently derived national estimates. A national survey esti mated the cumulative use in Austria to be between 2300 and 2800 t (Maderner and Hobiger, 1996), while our approach-which relies on GDP as a surrogate parameter-suggest a cumulative con sumption of 3075 t. In a similar survey for Norway, the cumulative consumption has been estimated to be approximately 1230 t (SFT, 1996), while our data suggest 944 t. 3.2.2. Export from OECD to other countries The amount of PCBs reported or estimated as being exported to non-OECD countries equals 148.3 kt or 11.2% of the total global production. We generally assumed that no PCBs were exported from OECD to China, the USSR or Eastern Europe. For the sake of simplicity, we further assumed no export to countries with a Gross Domestic Product (GDP) of less than US$ 1000 per capita or total GDP of less than one billion US$. The export to non-OECD countries includes 69 different countries. While the criteria to exclude some countries are somewhat arbitrary, these coun tries (i.e. approx. 70) would have accounted for a potential consumption of only 6.3% of the total export to non-OECD countries (or 0.7% of the total global production) -- if included. This sug gests that this simplification introduces minor uncertainties into the overall inventory at the scales of interest. 3.2.3. Eastern Europe According to Sabata et al. (1993), approximate ly half of the amounts of PCBs produced in LEXOLDMON003395 WATER_PCB-00000025 194 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 Czechoslovakia was exported to other Eastern European countries. Hence, 50% of the annual Czechoslovakian production of PCBs was distrib uted among the countries of Eastern Europe, according to the Gross Domestic Product of the countries within the region. For comparison, the total export from the Czechoslovakia adds up to 21.5 kt or 1.6% of total global production. 3.2.4. Countries within the former Soviet Union According to AMAP (2000), 60% of the Sovtol produced was used in Russia, and the rest in the former republics of the USSR. Similarly, 60% of the industrial capacitors containing PCBs were used in Russia. As a general assumption, we thus assumed that the former Soviet Union was a closed market where 60% of the total production was used in Russia. The remaining 40% were distrib uted among the other states of the former USSR according to Gross Domestic Product. Altogether, it is estimated that 173.8 kt or 13.1% of the global production have been used in the former Soviet Union. 3.2.5. China Only a minor quantity of PCBs is reported to have been produced in China (8000 t or 0.6% of the total global production). We assume that all of the PCBs produced has been used within China. Fig. 5 shows the estimated cumulative global consumption pattern for total PCBs, and includes estimates for 114 individual countries. It is esti mated that USA has been responsible for as much as approximately 46% of the total historical global PCB consumption. Other major consuming coun tries include Russia (7.9%), Germany (7.1%), Japan (4.1%), France (4.1%), Canada (3.0%), Ukraine (2.4%), Spain (2.4%), Italy (2.1%) and UK (2.0%). The data at the national level should be interpreted with great caution. Particularly for the countries relying entirely on the assumptions related to GDP. It is, however, imperative to keep in mind that the overall objective of this study is to try to capture the overall spatial pattern of PCB consumption at a global scale. If the emphasis was at the national level, a different approach would be required. In Fig. 5, the spatial distribution of the total historical national consumption is based on popu lation densities within each country by use of the GEIA grid system of 1 by 1 (Dr Yi-Fan Li, Environment Canada). Population density is con sidered a suitable surrogate parameter, as the con sumption of PCBs is generally linked with the use of electrical equipment. Overall, the results suggest that almost 97% of the intentionally produced PCBs have been used in the Northern Hemisphere. Furthermore, approximately 18% of the total have been used between 40 and 42 Northern latitude. The results also show that there are temporal changes in the latitudinal distribution of PCB consumption. Fig. 6 shows how the latitudinal consumption patterns changed in time for total PCBs. As can be seen from this figure, the highest peak occurred in the 1960s (Fig. 6D). In the last period considered, the figure reflects the continu ing consumption of PCBs in countries of the former Soviet Union (Fig. 6F). The resulting dataset reveals that there are tem poral and spatial variations in the homologue and congeneric consumption pattern. Fig. 7 exemplifies the regional differences in the historical consump tion pattern for selected PCB congeners. The regions were selected in a way that is reflecting the method used to estimate the global consump tion (see Fig. 4). We observe from Fig. 7 that PCB-101 and PCB-118 appear to be of limited significance in the consumption estimates for East ern Europe as compared to the other regions. Furthermore, PCB-180 has been of relatively minor importance among the selected congeners in the former Soviet Efnion and China. However, these differences should be interpreted with care. For example, according to Takasuga et al. (1996), the technical mixture produced in Poland resem bles Aroclor 1260. Hence, the estimates presented here for consumption in Poland are likely biased towards the lighter congeners as the production rates are lacking and hence not included in this inventory. The information may, however, be more reliable in other countries. For example, the avail able information would suggest that homologues with more than eight chlorines have not been used in the former Soviet Union (see also Table 2). 4. Final remarks The global spatial and temporal consumption pattern of individual PCBs is considered essential LEXOLDMON003396 WATER_PCB-00000026 K. Breivik et al. / The Science of the Total Environment 290 (2002) ]81-198 195 Fig. 5. Estimated cumulative global usage of PCBs (legends in t) with Px 1 longitude and latitude resolution. information for the interpretation of global PCB contamination patterns. Considering the large tem poral and spatial scales of this approach, it is difficult to ensure that all relevant information has been considered. Indeed, we expect that for certain countries and years, more reliable data are availa ble than those included in this study. Furthermore, only selected aspects of the involved uncertainties could be addressed here in a meaningful and quantitative way. We can presently only recognise other sources of uncertainty in a qualitative man ner. For example, we are certain that there has been some production of PCBs in other countries, and that the compiled production data may be underestimated. Another source of uncertainty is that the estimated production of individual homologues and congeners for the first decades essen tially remains unknown, but these data are obviously of less relative importance for current environmental levels. In spite of these uncertain ties, we are confident that important aspects of the temporal and spatial pattern of global consumption are reflected in this inventory, although the uncer tainties may be significant at a more detailed level, e.g. for single consuming countries, year and for some individual congeners. The availability of information for major producing companies and consuming countries around the time of their peak production indicates that the recent data are more accurate than the data from the past. 5. Additional information Selected data from this study are available through internet as Microsoft Excel spreadsheets at www.ni1 u.no/projects/globalpcb/. LEXOLDMON003397 WATER_PCB-00000027 196 K. Breivik et al. / The Science of the Total Environment 290 (2002) J8J-J98 Tonnes 40000 35000 (a) 1930-1939 40000 35000 Kb) 1940-1949 30000 30000 25000 20000 15000 25000 | 20000 H 15000 10000 5000 0 A 10000 5000 0 J-A -90 -60 -30 0 30 60 90 -90 -60 -30 0 30 60 90 Tonnes Tonnes Fig. 6. Global consumption of total PCBs for six different time-periods (by latitude). BPCB-28 PCB-52 PCB-101 PCB-118 B PCB-138 PCB-153 PCB-180 Fig. 7. Regional differences in the cumulative consumption of selected PCB congeners (in percent). LEXOLDMON003398 WATER_PCB-00000028 K. Breivik, et ai / The Science of the Total Environment 290 (2002) J8J-J9S 197 Acknowledgments We acknowledge financial support from the EU research project Global SOC (ENV4-CT97-0638). We are also grateful to other members of the Global SOC project for many stimulating discus sions. Technical assistance from Sverre Solberg and Finn Bjorklid in the preparation of the map is further appreciated. Finally, we are indebted to Dr Frank Wania for valuable comments and construc tive criticism on this paper. References Ahlborg UG, Hanberg A, Kenne K. Risk Assessment of Polychlorinated Biphenyls (PCBs). Nord 1992:26. ISBN 92 9120 075 1. Nordic Council of Ministers. Copenhagen, Denmark. 1992. (p. 99). AMAP. AMAP Assessment Report: Arctic Pollution Issues. Arctic Monitoring and Assessment Program (AMAP), Oslo, Norway, 1998, xii+ 859 pp. ISBN 82-7655-061-4. AMAP. PCB in the Russian Federation: Inventory and Propos als for Priority Remedial Actions. Executive Summary. AMAP Report 2000: 3, 2000. ISBN 82-7971-008-6. 27 pp. Ballschmiter K, Zell M. Analysis of polychlorinated biphenyls (PCB) by glass capillary gas chromatography. Composition of technical Aroclor- and Clophen-PCB mixtures. Fresenius Z Anal Chem 1980;302:20-31. Breivik K, Sweetman A, Pacyna JM, Jones KC. Towards a global historical emission inventory for selected PCB con geners -- a mass balance approach. 2 Emissions. Sci Total Environ 2002;290:167-184. Bruckmeier BFA, Jiittner 1, Schramm K-W, Winkler R, Stein berg CEW, Kettrup A. PCBs and PCDD/Fs in lake sedi ments of Grosser Arbersee, Bavarian Forest, South Germany, Environ Pollut 1997;95:19-25. Christensen ER, Lo C-K. Polychlorinated biphenyls in dated sediments of Milwaukee Harbour, Wisconsin, USA. Environ Pollut (Series B) 1986;12:217-232. Cummins JE. Extinction: the PCB threat to marine mammals. Ecologist 1988; 18(6): 193-195. de Voogt P, Brinkman UATh. Production, properties and usage of polychlorinated biphenyls. In: Kimbrough RD, Jensen AA, editors. Halogenated biphenyls, terphenyls, naphtalenes, dibenzodioxins and related products. Fopics in environmen tal health. ISBN 0-444-81029-3Elcscvicr, 1989:3-45. Edwards R. The polychlorobiphenyls, their occurrence and significance: a review. Chem Ind 1971;47:1340-1348. Falandysz J, Yamashita N, Tanabe S, Tatsukawa R. Composi tion of PCB isomers and congeners in technical Chlorofen formulation produced in Poland. Int J Environ Anal Chem 1992;47:129-136. Fiedler H. Polychlorinated biphenyls (PCBs): use and envi ronmental releases. Proceedings of the Subregional Meeting on Identification and Assessment of Releases of Persistent Organic Pollutants (POPs). St. Petersburg, Russian Federa tion, 1-4 July 1997. Inter-Organization Programme for the Sound Management of Chemicals 1997:81-103. Frame GM. A collaborative study of 209 PCB congeners and 6 Aroclors on 20 different HRGC columns. Part 2. Semiquantitative Aroclor congener distributions. Fresenius J Anal Chem 1997;357:714-722. Hamer T, Kylin H, Bidleman TF, Halsall C, Strachan WMJ, Barrie LA, Fellin P. Polychlorinated napthalenes and coplanar polychlorinated biphenyls in Arctic air. Environ Sci Technol 1998;32:3257-3265. Hillery BR, Girard JE, Schantz MM, Wise SA. Characterisa tion of three Aroclor mixtures using a new cyanobiphenyl stationary phase. Fresenius .1 Anal Chem 1997;357:723 -- 731. Ivanov V, Sandell E. Characterization of polychlorinated biphenyl isomers in sovol and trichlorodiphenyl formula tions by high-resolution gas chromatography with electron capture detection and high-resolution gas chromatography -- mass spectrometry techniques. Environ Sci Technol 1992;26:2012-2017. Iwata H, Tanabe S, Sakai N, Nishimura A, Talsukawa R. Geographical distribution of persistent organochlorines in air, water and sediments from Asia and Oceania, and their implications for global redistribution from lower latitudes. Environ Pollut 1994;85:15-33. Jensen S. Report of a new chemical hazard. New Scientist 1966;32:312. Jiang K, Li L, Chen Y, Jin J. Determination of PCDD/Fs and dioxin-like PCBs in Chinese commercial PCBs and emis sions from a testing PCB incinerator. Chemosphere 1997;34:941-950. Kalmaz EV. Kalmaz GD. Transport, distribution and toxic effects of polychlorinated biphenyls in ecosystems: review. Ecol Model 1979;6:223-251. Kannan N, Schulz-Bull DE, Petrick G, Duinker JC. High resolution PCB analysis of Kanechlor, Phenoclor and Sovol mixtures using multidimensional gas chromatography. Int J Environ Anal Chem 1992;47:201-215. Maderner W, Hobiger G. PCB-Stoffbilanz-Osterreich (in German) Zusammenfassung (Abstract only). Monographie, Band 79. Umweltbundesamt, Wien, 1996. Oehme M, Mano S. The long-range transport of organic pollutants to the Arctic. Fresenius Z Anal Chem 1984;319:141-146. Oehme M. Further evidence for long-range air transport of polychlorinated aromates and pesticides: North America and Eurasia to the Arctic. Ambio 1991 ;20(7):293--297. Rapaport RA, Eisenreich SJ. Historical atmospheric inputs of high molecular weight chlorinated hydrocarbons to eastern North America. Environ Sci Technol 1988;22:931-941. Sabata S, Friesova A, Rericha R, Hetflejs J. Limits to the use of KOH/PEG method for destruction of PCB liquids of LEXOLDMON003399 WATER_PCB-00000029 198 K. Breivik et al. / The Science of the Total Environment 290 (2002) 181-198 Czechoslovak production. Chemospherc 1993:27(7): 1201 1210. Sanders G, Jones KC, Hamilton-Taylor J, Dorr H. Historical inputs of polychlorinated biphenyls and other organochlorines to a dated lacustrine sediment core in rural England. Environ Sci Technol 1992;26:1815-1821. Schlosserova J. 'Kontrolle ausgewahlter Boden in der Tschechiscben und Slowakischen Republik auf ihre Kontamination mit chlorierten Kohlenwassertstoffen (In German). In: Heinisch E, Kettmp A, Wenzel-Klein S, editors. Schadstoffatlas Osteuropa. ISBN 3-609-69540-4 Germany: Ecomed Verlagsgesellschaft AG & Co, 1994:54-59. Schulz DE, Petrick G, Duinker JC. Complete characterization of polychlorinated biphenyl congeners in commercial Aroclor and Clophen mixtures by multidimensional gas chro matography -- electron capture detection. Environ Sci Technol 1989;23:852-859. SFT, PCB i Norge-Forekomsl og forslag til tiltak. (In Nor wegian). Norwegian Pollution Control Authority, Oslo, Nor way, 1996. Report 96:08. 51 pp. ISBN 82-7655-352-4. Takasuga T, Inoue T, Ishida T, Ireland P. Determination of the composition of the commercial PCBs, Kanechlor, Clophen, Aroclor, Chlorofen, and Sovol, by HRGC-HRMS. Organohalogen Compound 1996;27:391-396. J'anabe S. PCB problems in the future: foresight from current knowledge. Environ Pollut 1988;50:5-28. Tatsukawa R. PCB pollution of the Japanese environment. In: Higuchi K, editor. PCB poisoning and pollution. ISBN 0 12-347850-2 Kodansha Ltd. Tokyo: Academic Press, 1976:147-179. United Nations. Statistical yearbook. Thirty-ninth issue. New York, USA. ISBN 92-1-061159-4 1994. Vallack HW, Bakker DJ, Brandt I, Brorstrom-Lunden E, Brouwer A, Bull KR, Gough C, Guardans R, Holoubek 1, Jansson B, Koch R, Kuylenstierna J, Lecloux A, Mackay D, McCutcheon P, Mocarelli P, Taalman RDF. Controlling persistent organic pollutants -- what next? Environ Toxicol Pharmacol 1998;6:143-175. Voldner EC, Li Y-F. Global usage of selected persistent organochlorines. Sci Total Environ 1995:160/161:201-210. Waid JS, editor. PCBs and the environment. Volume I, 2 and 3. ISBN 0-8493-5923-6 (V.l) 0-8493-5924-4 (V.2) and 0 8493-5925-2 (V.3) Boca Raton, Florida: CRC Press, Inc, 1986. Wania F, Mackay D. Tracking the distribution of persistent organic pollutants. Environ Sci Technol 1996;30(9):390A396. Wh 0. Polychlorinated biphenyls and terphenyls. Second Edition. Environmental Health Criteria 140 Geneva: World Health Organization. ISBN 92 4 157140 3, 1993. Xu S, Jiang X, Dong Y, Sun C, Feng J, Wang L, Martens D, Gawlik BM. Polychlorinated organic compounds in Yangtse River sediments. Chemosphere 2000;41:1897-1903. LEXOLDMON003400 WATER_PCB-00000030