AREA 6 • MODELING • RESEARCH ARTICLE

Long-range atmospheric transport of three toxaphenecongeners across Europe. Modeling by chained single-boxFATEMOD program

Jaakko Paasivirta & Seija Sinkkonen &

Vladimir Nikiforov & Fedor Kryuchkov &

Erkki Kolehmainen & Katri Laihia & Arto Valkonen &

Manu Lahtinen

Received: 6 September 2008 /Accepted: 27 October 2008 / Published online: 9 January 2009# Springer-Verlag 2008

AbstractBackground, aims, and scope Since toxaphene (polychloro-camphene, polychloropinene, or strobane) mixtures wereapplied for massive insecticide use in the 1960s to replacethe use of DDT, some of their congeners have been found athigh latitudes far away from the usage areas. Especiallypolychlorinated bornanes have demonstrated dominatingcongeners transported by air up to the Arctic areas.Environmental fate modeling has been applied to monitorthis phenomenon using parallel zones of atmosphere aroundthe globe as interconnected environments. These zones,shown in many meteorological maps, however, may not bethe best way to configure atmospheric transport in air

trajectories. The latter could also be covered by connectinga chain of simple model boxes. We aim to study thisalternative approach by modeling the trajectory chain usingcatchment boxes of our FATEMOD model. Polychlorobor-nanes analyzed in biota of the Barents Sea offered one caseto study this modeling alternative, while toxaphene hasbeen and partly still is used massively at southern EastEurope and around rivers flowing to the Aral Sea.Materials and methods Pure model substances of threepolychlorobornanes (toxaphene congeners P26, P50, andP62) were synthesized, their environmentally importantthermal properties measured by differential scanningcalorimetry, as evaluated from literature data, and theirtemperature dependences estimated by the QSPR programsVPLEST, WATSOLU, and TDLKOW. The evaluatedproperty parameters were used to model their atmosphericlong-range transport from toxaphene heavy usage areas inUkraine and Aral/SyrDarja/AmuDarja region areas, throughEast Europe and Northern Norway (Finnmarken) to theBarents Sea. The time period used for the emission modelwas June 1997. Usual weather conditions in June wereapplied in the model, which was constructed by chainingFATEMOD model boxes of the catchment’s areas alongassumed maximal air flow trajectories. Analysis of the threechlorobornanes in toxaphene mixtures function as a basisfor the estimates of emission levels caused by its usage.High estimate (A) was taken from contents in a Westernproduct chlorocamphene and low estimate (B) from meancontents in Russian polychloroterpene products to achievemodeled water concentrations. Bioaccumulation to ana-lyzed lipid of aquatic biota at the target region wasestimated by using statistical calculation for persistentorganic pollutants in literature.

Environ Sci Pollut Res (2009) 16:191–205DOI 10.1007/s11356-008-0084-2

This article is dedicated to the memory of Professor Alexander BTerentiev (who passed away in November 2006), our true friend. Withhis Institute of Organo-Element Compounds, Russian Academy ofScience, Moscow, he was an important main organizer of the six jointFinnish–Russian seminars (every third year since 1989) on the field(‘Chemistry and Ecology of Organo-Element Compounds’). Heprompted us especially to search properties and environmental fatesfor various polyhalogen compounds. We remember him for hisfriendly character and great sense of humor.

Responsible editor: Cafer Turgut

J. Paasivirta (*) : S. Sinkkonen : E. Kolehmainen :K. Laihia :A. Valkonen :M. LahtinenDepartment of Chemistry, University of Jyväskylä,P.O. Box 35, 40014 Jyväskylä, Finlande-mail: jpta@jyu.fi

V. Nikiforov : F. KryuchkovDepartment of Chemistry, St. Petersburg State University,Universitetskii pr. 26,198504 St. Petersburg, Russia

Results The results from model runs A and B (high and lowemission estimate) for levels in sea biota were compared toanalysis results of samples taken in August 1997 at BarentsSea. The model results (ng g−1 lw): 4–95 in lipid ofplanktovores and 7–150 in lipid of piscivores, were in fairagreement with the analysis results from August 1997: 21–31 in Themisto libellula (chatka), 26–42 in Boreocadussaida (Polar cod), and 5–27 in Gadus morhua (cod) liver.Discussion The modeling results indicate that the applica-tion of chained simple multimedia catchment boxes onpredicted trajectory is a useful method for estimation ofvolatile airborne persistent chemical exposures to biota inremote areas. For hazard assessment of these pollutants,their properties, especially temperature dependences, mustbe estimated by a reasonable accuracy. That can beachieved by using measurements in laboratory with puremodel compounds and estimation of properties by thermo-dynamic QSPR methods. The property parameters can bevalidated by comparing their values at an environmentaltemperature range with measured or QSPR-estimatedvalues derived by independent methods. The chained boxmethod used for long-range air transport modeling can bemore suitable than global parallel zones modeling usedearlier, provided that the main airflow trajectories andproperties of transported pollutants are predictable enough.Conclusions Long-range air transport modeling of persis-tent, especially photo-resistant organic compounds using achain of joint simple boxes of catchment’s environments isa feasible method to predict concentrations of pollutants atthe target area. This is justified from model resultscompared with analytical measurements in Barents Seabiota in August 1997: three of six modeled values werehigh and the other three low compared to the analysisresults. The order of magnitude level was similar in bothmodeled (planktovore and piscivore) and observed (chatkaand polar cod) values of lipid samples. The obtained resultswere too limited to firm validation but are sufficient tojustify feasibility of the method, which prompts one toperform more studies on this modeling system.Recommendations and perspectives For assessment of therisk of environmental damages, chemical fate determinationis an essential tool for chemical control, e.g., for EUfollowing the REACH rules. The present conclusion ofapplicability of the chained single-box multimedia model-ing can be validated by further studies using analyses ofemissions and target biota in various other cases. Toachieve useful results, fate models built with databaseshaving automatic steps for most calculations and outputsaccessible to all chemical control professionals are essen-tial. Our FATEMOD program catchments at environmentsand compound properties listed in the database represent afeasible tool for local, regional, and, according our presenttest results, for global exposure predictions. As an extended

use of model, emission estimates can be achieved byreversed modeling from analysis results of samplescorresponding to the target area.

Keywords Advection . Catchment’s areas .

Chain of model boxes . Chlorobornanes .

Degradation half-life times . Degradation rates . DSC .

Emission . FATEMOD . Joint trajectories . Photolysis .

Temperature coefficients

1 Background, aim, and scope

Some persistent organic pollutants (POPs) have beenobserved to accumulate in Arctic and Antarctic biota,despite the fact that their exploitation and/or spillage takeplace only at lower latitudes. Few congeners of the widelyused chlorohydrocarbon insecticides, such as toxaphenemixtures (Saleh 1991; Vetter and Oehme 2000), are aspecial case of such hazard (AMAP 2002). Soon afterstarting the massive usage of these compounds, gaschromatographic analyses started to show pattern oftoxaphene peaks in air, water sediment, and biota samples(Tabor 1966; Nicholson 1968; Nicholson et al. 1968).Long-range air transport of toxaphene (like several otherorganochlorine pesticides) up to the Arctic marine foodchains was detected to be an important ecological hazard(Bidleman and Olney 1975; Norstrom et al. 1988).Toxaphene residues have even been stated to form anoverall global pollution hazard perhaps larger than PCB(Zell and Ballschmiter 1980). This could be understood byextraordinary high volatilities of the toxaphene congeners:high Henry’s law coefficient=vapor pressure/water solu-bility (Murphy et al. 1987). Very low water solubility andhigh persistency of some major toxaphene components inair, especially chlorobornanes, explain their occurrence inthe Arctic (Barrie et al. 1993). In Sweden, toxapheneprofile was detected by GC–ECD profile in rainwater(Sundström 1981), but (by GC–MS) only at very lowlevels in mother’s milk in Sweden and Finland (Bidlemanet al. 1987; Vaz and Blomqvist 1985). Instead, levels inburbot liver were up to several μg g−1 in lipid (Pyysalo andAntervo 1985) and levels in Arctic salmon and cod liverwere from some μg g−1 up to 126 μg g−1 in lipid (Tarhanenet al. 1989; Paasivirta and Rantio 1991; Paasivirta et al.1993).

While toxaphene mixtures were manufactured by chlo-rination of camphene, pinene, or other monoterpenes, theirtoxic components were suspected to be formed also inchlorobleaching of pulp and, thus, contribute to hazard ofeffluents of paper industries, e.g., in the USA, Canada,Sweden, and Finland (Swackhamer et al. 1993). However,analyses by sensitive GC–MS methods (Rantio et al. 1993)

192 Environ Sci Pollut Res (2009) 16:191–205

showed many chlorinated terpenes but no toxaphenecongeners in effluent or biota at recipient waters close tothose mills. This can be explained by the hot cookingprocess in pulping which causes nearly total evaporation ofmonoterpenes before the residual pulp is bleached. Instead,fish from more remote water areas contained traces of sometoxaphene congeners. These contaminants obviously orig-inated from atmospheric long-range transport and deposi-tion by wet and solid fallout, like that observed, e.g., byanalyses of rainwater and atmosphere in Sweden, and bystudies on mother’s milk and wildlife in Baltic and Arcticareas (see references above).

Total global usage of toxaphene was estimated to be1.33 million tons from 1950 to 1993 and 0.67 million tonsfrom 1970 to 1993 (Voldner and Li 1993, 1995). Rapidresistance development in pests led to increased use, whichfurther caused catastrophes, e.g., to the agriculture in Egypt(El-Sebae et al. 1993). After a stop of all uses in USA in1986 (Saleh 1991), large amounts of toxaphene residuesremained in soil at both usage and non-usage areas andcould be analyzed also in air of the areas (Jantunen et al.2000; Harner et al. 2001; Li et al. 2001). The heavy usageareas certainly remain as significant sources of toxaphene toatmosphere tens of years after the end of use. One of theEuropean concerns are particularly the massive usage oftoxaphene and the similar mixture of polychloropinene insouthern states of former Soviet Union in sugar beet fieldsof Ukraine and cotton production areas in Uzbekistan,North-East Turkmenistan, Kazakhstan and Tadzhikistan(Fedorov 1999).

The observed environmental fate of toxaphene as aglobal pollutant was first supported by modeling itsatmospheric long-range transport from measured propertiesof the pesticide mixture. Firstly, Global Zonal fugacity massbalance models showed its feasibility to such evaluation(Wania and Mackay 1993a, b). The ‘weathering’ in themixture composition (McConnell et al. 1993) could betaken into account by measures of soil–air and water–airpartition at different climatic zones or boxes included inmodel trajectories (Wania and Mackay 1993a, b; McConnellet al. 1993; Swackhamer et al. 1999; MacLeod et al. 2002;Jantunen and Bidleman 2003; Bidleman and Leone 2004).

In 1980s, congener-specific analyses of toxaphenecomponents became possible, when individual compoundswere isolated from highly contaminated animals and theirstructures together with analytical properties were deter-mined (Vetter and Oehme 2000). Also, syntheses ofindividual pure model compounds were performed andtheir structures validated. Consequently, modeling environ-mental fate of individual congeners also became possible.Previous atmospheric transport models of toxaphene werebased on the properties of mixtures. However, if crystalsample (5–10 mg) of pure congener was available, reliable

values for enthalpy of fusion (ΔHf, J mol−1) and meltingpoint (Tm, K) could be measured by differential scanningcalorimetric methods (DSC). Consequently, an essentialparameter, entropy of fusion, ΔSf, can be calculated asΔHf /Tm (J K−1 mol−1) for temperature dependence factorsof water solubility (SW), Henry’s law function (H), andLogKOW. These parameters are useful in modeling fate ofindividual congeners in different environments and seasons(Paasivirta 2006). If a sufficient number (>10) of purecongeners would have been available for DSC, the ΔSfvalues could be estimated for all substances in a congenergroup using multiple linear regression of proper moleculardescriptors. For example, determination of reliable SW, H,and LogKOW values for all 75 polychlorinated naphthalenes(PCNs) became possible by DSC analyses of 11 purecongeners (Lahtinen et al. 2006).

The aim of this study was to estimate environmentallyrelevant properties, including temperature correction coef-ficients (Paasivirta 2006), of three major polychlorobornanesfound globally in fish, for use in the chains of single-boxmodels (FATEMOD) for their long-range transport fromUkraine and Aral Sea areas in typical summertimeconditions. Because only three compounds were involvedin this study, they were prepared to an outmost purity fordetermining their modeling parameters as accurately aspossible. This was necessary to elucidate the applicabilityof this modeling alternative.

2 Materials and methods

Nomenclature of polychlorobornanes IUPAC nomenclaturesuggested by Andrews and Vetter (1995) together with“AV” codes is recommended, because it specifies thestructures of both enantiomers and allows new structuresfound to be added to their list without confusion. Theformula of the three congeners abundant in Barents Sea fishand their names and codes are illustrated in Fig. 1. Becauseonly three congeners, abundant in the Arctic, are consid-ered, codes of analytical standard compounds from Parlar etal. (1995) will be used in text and tables for simplicity.

Preparation of the model compounds First, analyticallypure polychlorobornane samples were isolated from tech-nical toxaphene (Matsumura et al. 1975; Turner et al. 1977;Chandurkar et al. 1978; Swanson et al. 1978; Chandurkarand Matsumura 1979). In the beginning of the 1990s, firstcongeners were isolated from Beluga whales and structuresdetermined by NMR and mass spectrometry (Stern et al.1992; Burhenne et al. 1993), methods that have remainedsince then as tools for identification of known congenersand structure determination of previously unknown con-geners. Studies of compounds isolated from biota were

Environ Sci Pollut Res (2009) 16:191–205 193

supported by syntheses: chlorinations of monoterpenes andisolation of individual congeners by preparative chroma-tography (Burhenne et al. 1993; Kallenborn et al. 1994;Tribulovich et al. 1994). The nomenclature of the struc-tures, however, deviated from the later reverse numberingof the bridge carbons, which was brought into play in 1995according to the IUPAC rules (C8 at the side of ringcarbons 5–6, C9 at the side of ring carbons 2–3). Thiscaused confusion, especially in the names of enantiomers.Configurations and correct, now generally adopted (Vetterand Oehme 2000; Andrews and Vetter 1995) names andcodes of the compounds P26, P50, and P62 are shown inFig. 1.

Pure samples of the chlorobornanes used in this studywere prepared from the common intermediate—2-exo,10,10-trichlorobornane (Tribulovich et al. 1994). Thisintermediate was obtained by portion-wise addition ofchlorine to a solution of camphene in carbon tetrachlorideuntil green color remained for several minutes. The solventwas evaporated off and distillation of residue in vacuoyielded crude 2-exo,10,10-trichlorobornane, which wasused without further purification. For preparation of P26,2-exo,10,10-trichlorobornane in carbon tetrachloride waschlorinated further under a UV lamp until content of P26 inthe reaction mixture reached approximately 1% (GC–ECD).

This corresponded to an average of eight Cl atoms permolecule of polychloroterpene. This mixture was separatedon a large glass column (165 сm high, 4.4 cm i.d.) filledwith silica gel, with hexane as an eluent. A fraction from8.4 to 8.6 L contained P26 as the major component. Afterevaporation of the major part of hexane, the productdeposited as a white solid. Several successive crystalliza-tions from hexane yielded pure P26.

For preparation of P50 and P62, 2-exo,10,10-trichlor-obornane in carbon tetrachloride was exhaustively chlori-nated under a UV lamp until the reaction stopped and themixture of nona- and decachlorobornanes was formed(GC–ECD). This mixture was separated in a similar manneras for the preparation of P26. A fraction from 12.7 to13.0 L contained P50 and a fraction from 13.8 to 14.2 Lcontained P62 as the major components. Evaporation of themain part of solvent from these fractions followed byseveral crystallizations yielded pure P50 and P62. The threecongeners were characterized by their relative retentiontimes in GC of polychloroterpene mixtures (Nikiforov et al.2000a, b). Structure proof of these congeners was obtainedby single crystal X-ray diffraction measurements whichgave the same outputs as published by Frenzen et al.(1994). Also, their NMR chemical shifts were identicalwith those published by Burhenne et al. (1993). Addition-ally, we performed correlation NMR studies of thesecongeners to find out whether the spin coupling situationswere in agreement with the structures.

Thermal analyses of the model polychlorobornanes P26,P50, and P62 Thermal transitions of the toxaphenecongeners were determined on a power compensation typePerkin-Elmer PYRIS DIAMOND DSC device. The meas-urements were carried out under nitrogen atmosphere (flowrate 50 mL min−1) using 50 μL sealed aluminum samplepans. The sealing was made by using a 30-μL aluminumpan with capillary holes as a cover piece to ascertain goodthermal contact between a sample and pan, and to minimizethe free volume inside the pan. The temperature calibrationwas made using two standard materials (n-decane, In) andenergy calibration by an indium standard (28.45 J g−1).Typically, the sample was heated from 25 to ~20–30°Cabove the predetermined melting transition with a heatingrate of 10°C/min, held 1 min at turning point, then cooleddown to −40°C and held 5 min before the heating scan wasrepeated. In addition, some measurements were made foreach compound, in which start of decomposition wassearched out by heating the samples to ~250–270°C.Sample weights of 0.5–5 mg were used on the measure-ments and each sample weight was checked afterwards tomonitor weight loss that may have occurred during thescans. The uncertainty for measured temperatures was lessthan 0.7°C for all measurements. The reported values are

Fig. 1 Structures and nomenclature of the three abundant polychlor-obornanes in Barents Sea fish. Names below the formula areaccording to IUPAC rules (Andrews and Vetter 1995). B8/B9 symbolsare systematic codes of Andrews and Vetter (1995), where a or bextensions mean different enantiomers. P symbols are from Parlar’sreference standard mixture (Parlar et al. 1995)

194 Environ Sci Pollut Res (2009) 16:191–205

obtained as a mean of the two analogous measurements,and all the melting points were taken at the peak maximum(Table 1; see also Fig. 3).

FATEMOD model and linked single-box models for long-range transport estimation FATEMOD (Paasivirta et al.2002; Paasivirta 2006) is a single-box multimedia modelwhere fate of a chemical in constant emissions andadvections can be modeled for steady state, a timescaleafter stop of emission, and bioaccumulation estimates ofPOPs to four trophic levels of aquatic species. It is based onthe fugacity modeling algorithms (GENERIC) of Mackay(1991), and on additional programs which utilize catchmentareas as environment boxes. Utilizing environment andsubstance databases and automatic temperature and pHcorrections ON LINE to properties of chemical make theuse of FATEMOD practical for purposes of risk estimationsof chemicals, e.g., in REACH activities. Although theFATEMOD is a rather simple method, its rational use aspart of multi-box model systems is not ruled out. Rathersimilar boxes have previously been linked to the fatemodeling net in the POPCYCLING BALTIC (Pacyna et al.1999; Wania et al. 2000). In the present study, we linkFATEMOD boxes to a chain for modeling atmospherictransport of the three very persistent polychlorobornanesfrom areas of the former massive toxaphene usage towardsthe northwest where their traces have been analyzed inbiota of the Barents Sea. The model of this catchment areachain does not contain the less or non-connected areas ofthe zonal multi-box models (Wania and Mackay 1993a, b;Wania et al. 2000). More sophisticated atmospherictransport models for fate of POPs have been recentlypresented and compared to previous models (Semeena andLammel 2003; Lammel et al. 2007).

Properties of environments Properties of the model envi-ronments for FATEMOD database are mainly obtainedfrom scientific literature. The environments for mostcommon regional modeling are selected as catchment areas.For each country, the GLOBACK data, available together

with CLOBOX model (Sleeswijk 2006), were used as aprincipal source of properties. The FATEMOD modelenvironment consists of four media compartments (i): Air(1), Water (2), Soil/Plants (3), and Sediment (4). The twosub-compartments are Suspended Solids (5) and WaterBiota or ‘Fish’ (6). Values of compartments in environ-mental database are Area (AR, m2), Height (HT, m), Volume(V, m3; includes sub-compartments), fraction of organiccarbon (ORG), advective residence time (GRA, h), fractionof solids in water (FSSW, default 0.00001), fat fraction infish (FFF, default 0.05), rain rate (U3, m h−1), emission rate(EK), pH of surface water (WPH), and pH of groundwater(SWPH). Values of these parameters can be changed ONLINE via the ‘edit environments’ mode. The otherenvironmental conditions can be changed only by editingthe program code itself. Environments (catchment areas)and their properties in FATEMOD database are listed inTable 2. A map of the areas of the FATEMOD chain isshown in Fig. 2.

Constant properties of chemicals for FATEMOD Propertiesof chemicals (Table 3) are obtained partly from literature,but in great part by theoretical, statistical, and structuralparameter calculation (QSPR) methods (Mackay 1991;Pacyna et al. 1999; Paasivirta et al. 2002; Paasivirta2006). Basic constants needed for model property estima-tion are molecular mass (WM, g mol−1), melting point (Mp,°C→TM, K), entropy of fusion (ΔSf, J K−1 mol−1), molarvolume (Vb, cm3 mol−1; Ruelle et al. 1991), and solubilityparameter (DB, MPa1/2; Ruelle 2000). Associating groups,proton acceptors, and donors are counted from the structureand used as additional incremental parameters in theevaluation of SW and LogKOW. In alicyclic toxaphenecongeners, each chlorine substituent was estimated fromRuelle (2000) or our tests to determine an increment inwater KaccW=25 for octachlorobornane P26 and 26.5 fornonachlorobornanes P50 and P52. Values of pHs in modelbox waters are considered in FATEMOD by automaticcorrection (Trapp and Matthies 1998) using pKa (forproperty of acidic or basic substrate only).

Table 1 Thermal transitions, enthalpies, and decomposition temperatures of P26, P50, and P62

Comp. WM 1st and 2nd heatinga Dec.Tm (DH) or Tg [DCp]; °C (J g−1); °C [J g−1 °C−1] Td (°C)

1 P26 413.82 Tm 154.1 (59.543) 231Tc-m-c, Tm 149.1 (30.88)

2 P50 448.26 Tm 132.9 (46.863) 245Tg −13.9 [0.11]

3 P62 448.26 Tm 148.6 (48.223) 205Tg −20.1 [0.45]

Tm melting transition, Tg glass transition, Td decompositiona Transitions from second heating scans are presented on the second row, and are taken from scans having end temperature of 180°C

Environ Sci Pollut Res (2009) 16:191–205 195

Temperature-dependent properties for FATEMOD QSPRalgorithms for estimates of the temperature correctioncoefficients (Ai, Bi, Eq. (1)) have been developed byPaasivirta et al. (1999; 2002) and Paasivirta (2006) asalgorithms VPLEST, WATSOLU, TDLKOW, and a methodto estimate coefficients for the Henry’s law function.

LogPropertyi ¼ Ai� Bi=T Kð Þ ð1Þ

The Eq. (1) type to make correction to the physical propertyof chemicals was first discovered by Clausius and Cla-peyron in the 1850s for vapor pressure PS or PL (in solid or(subcooled) liquid state, respectively). The similar Ai andBi coefficients have been much more recently found usefulfor adjustment of the values of other properties like watersolubility SW, octanol–water partition KOW, and Henry’slaw function H. Applications of Ai and Bi corrections arelisted as Eqs. (2)–(10).

VPLEST LogPS ¼ Aps� Bps=T ð2Þ

LogPL ¼ Apl� Bpl=T ð3Þ

Conversions Apl ¼ Aps�ΔSf=19:1444 ð4Þ

Bpl ¼Bps�ΔSf*TM=19:1444 ð5Þ

WATSOLU LogSW ¼ As� Bs=T ð6Þ

TDLKOW LogKOW ¼ Aow� Bow=T ð7Þ

Henry's function LogH ¼ Ah� Bh=T ð8Þ

Ah ¼ Apl� As ð9Þ

Bh ¼ Bpl� Bs ð10Þ

In environment, organic chemical is distributed to gas,solvated molecules, or ions. Therefore, the only vaporpressure value needed in model calculations is PL. If only

Table 2 Properties of environments for FATEMOD database

Environmenta AR(1) AR(2) AR(3) AR(4) HT(1)

HT(2)

HT(3)

HT(4)

ORG(3)

ORG(4)

ORG(SS)

Source 1 (Ukraine fields) 6.03E+11 6.58E+09 5.96E+11 6.58E+09 9,450 6.2 0.1 0.02 0.02 0.04 0.20NWB (NW Russia/Belorussia) 3.50E+11 2.20E+08 3.50E+11 2.20E+08 9,300 2 0.1 0.02 0.02 0.04 0.20Source 2 (cotton fields) 1.04E+12 2.03E+09 1.04E+12 2.03E+09 9,600 5.5 0.1 0.02 0.02 0.04 0.20KAZ (West Kazakhstan) 1.45E+12 1.10E+10 1.44E+12 1.10E+10 9,500 7.2 0.1 0.02 0.02 0.04 0.20CWRU (Central West Russia) 1.68E+12 1.63E+11 1.52E+12 1.63E+11 9,400 12 0.1 0.02 0.04 0.05 0.12NEVA (Neva watershed) 3.71E+11 5.86E+10 3.12E+11 5.86E+10 9,200 60 0.1 0.02 0.05 0.04 0.10CFI (North Central Finland) 6.30E+10 4.42E+09 5.85E+10 4.42E+09 8,800 2 0.1 0.02 0.05 0.06 0.07KemR (Kemijoki River catch) 5.11E+10 2.20E+09 4.89E+10 2.20E+09 8,600 2 0.1 0.02 0.05 0.06 0.07F/L (Finnmark/Lappland) 5.00E+10 7.00E+09 4.30E+10 7.00E+09 8,400 4 0.1 0.02 0.04 0.05 0.07Barents Sea 1.20E+12 1.20E+12 1.00E+08 1.20E+12 8,100 200 0.1 0.02 0.03 0.01 0.02

Environmenta V(1) V(2) V(3) V(4) V(SS) b U3 m h−1 GRA(1)

GRA(2) GRA(3)

GRA(4)

Source 1 (Ukraine fields) 5.70E+15 4.08E+10 5.96E+10 1.32E+08 2.04E+05 6.45E−05 100 1.3E+03 1.E+07 2.E+06NWB (NW Russia/Belorussia) 3.26E+15 4.40E+08 3.50E+10 4.40E+06 2.20E+03 7.05E−05 100 2.5E+02 1.E+07 2.E+06Source 2 (cotton fields) 9.98E+15 1.12E+10 1.04E+11 4.06E+07 5.58E+04 2.35E−05 100 4.0E+03 1.E+07 2.E+06KAZ (West Kazakhstan) 1.38E+16 7.92E+10 1.44E+11 2.20E+08 3.96E+05 2.85E−05 100 4.0E+04 1.E+07 2.E+06CWRU (Central West Russia) 1.58E+16 1.96E+12 1.52E+11 3.26E+09 9.78E+06 4.11E−05 100 4.5E+04 1.E+07 2.E+06NEVA (Neva watershed) 3.41E+15 3.52E+12 3.12E+10 1.17E+09 1.76E+07 5.14E−05 100 4.0E+05 1.E+07 2.E+06CFI (North Central Finland) 5.54E+14 8.85E+09 5.85E+09 8.85E+07 4.42E+04 4.91E−05 100 3.0E+03 1.E+07 2.E+06KemR (Kemijoki River catch) 4.40E+14 4.40E+09 4.89E+09 4.40E+07 2.20E+04 4.00E−05 100 1.3E+04 1.E+07 2.E+06F/L (Finnmark/Lappland) 4.20E+14 2.80E+10 4.30E+09 1.40E+08 1.40E+05 3.42E−05 100 2.0E+04 1.E+07 2.E+06Barents Sea 9.72E+15 2.40E+14 1.00E+07 2.40E+10 1.20E+09 5.48E−05 100 1.0E+06 1.E+07 2.E+06

For abbreviations AR, ORG, HT, V, WPH, and SWPH see text. U3=rain rate m h−1 (annual U3×8,760 m a−1 ). GRA(i)=advection residencetimes h=volume m3 /(res. rate m3 s−1 ×3,600). Fraction of fish in water FFW=1.00E−06. Fat fraction of fish (default): FFF=0.05a Indexes of compartments (i): Air (1), Water (2), Soil/Plants (3), Sediment (4)b Volume fraction of suspended solids (SS) in water FSSW=5E−06

196 Environ Sci Pollut Res (2009) 16:191–205

PS value is known, conversion Eqs. (4) and (5) are used toobtain the values of PL (Paasivirta et al. 1999, 2002).VPLEST algorithm (Paasivirta et al. 2002) based on thevapor pressure determination equation of Grain (1990) wasused for estimation of Apl and Bpl for the threepolychlorbornanes. For input value, only one known vaporpressure at one temperature is needed. Using structureparameters from Grain’s tables, PL value for each temper-ature at a selected range (e.g., −2 to +30°C) in one-degreeintervals was calculated and then linear regression for Apl(intercept) and Bpl (−slope) was performed according toEq. (3). For validation, results were compared in the 0–30°Crange with those from indirect GC-based estimation of Apland Bpl (Bidleman et al. 2003).

Estimation method for SW was developed from mobileorder and disorder thermodynamic (MOD) equations forcalculating the values of SW25 (mol m−3) by Ruelle et al.(1991), Ruelle and Kesselring (1997), and Ruelle (2000).These equations were modified by us to the WATSOLUalgorithm (Paasivirta et al. 1999, 2002; Paasivirta 2006),grouping the terms without T for calculation of the constantAs and other equations with T to calculate negativeintercept Bs in the Clausius–Clapeyron type of Eq. (6).The resulting SW (as Log) is a single value, at least forhydrophobic molecules, corresponding to the liquid state

(Schwarzenbach et al. 1993; Ruelle and Kesselring 1997;Trapp and Matthies 1998; Paasivirta et al. 1999, 2002).There is no “solid state solubility” for such compounds inthe environment, while they are distributed to (hydrated)single molecules at low concentrations there. Micelles, etc.are another matter.

In analogy with the procedure for temperature correctionof SW above, we have modified the equations for KOW25from Ruelle (2000) to a calculation scheme for coefficientsAow and Bow (TDLKOW=Paasivirta et al. 1999, 2002;Paasivirta 2006). A number of new constant parametersderived by Ruelle were included in the computing code.The final resulting equation is (7).

In FATEMOD modeling, Ah and Bh coefficients are notnecessary to obtain a volatility value of H in modeltemperature. For validations, these values for Eq. (8) arestored to the database anyhow. FATEMOD automaticallyuses the known approximation H=PL/SW at the modeltemperature, which is real in case of hydrophobic chem-icals. But also Eq. (8) can be used, because Ah and Bh canbe readily converted from the known coefficients for PL

and SW (Eqs. (9) and (10)).Reaction half-life times were evaluated from the mea-

sured or evaluated rates of photolysis, chemical degrada-tion, and biodegradation at some reference temperature,

Fig. 2 Map of the model areas.Trajectories of the assumedgreatest concentrations ofchlorobornanes in air fromsources 1 and 2 to the BarentsSea

Environ Sci Pollut Res (2009) 16:191–205 197

most often at 25°C. Temperature corrections are doneinstantly by FATEMOD (Sinkkonen and Paasivirta 2000).Reference values of the degradation half-life times ofpolychlorobornanes were determined mainly from theirenergy data (Vetter and Scherer 1998; Heimstad et al. 2001;Coelhan and Maurer 2005).

Summary of the substance properties of the P26, P50,and P62 for FATEMOD is collected in Table 3. The mostdifficult parameter for estimates of half-life times was thebiodegradation rate k(b) h−1, a major process in soil/plantsand sediments, but only a minor process in water. It hasbeen observed that chlorobornanes P26, P50, and P62degrade faster than most other known chlorobornanes inanaerobic conditions (Vetter and Oehme 2000). This wastaken in account by increasing their biodegradation rates tosomewhat higher values related to a nominal valueobtainable from their formation energy. In the atmosphere,this degradation is practically absent. In addition, photo-

stability of these congeners is very high (Vetter and Scherer1998; Heimstad et al. 2001; Coelhan and Maurer 2005).

3 Results

The thermal behavior of model compounds Due to smallquantities of the toxaphene compounds available, theinitiation of decomposition of the samples was not observedby thermo-gravimetric analyzer but by DSC. All the resultsare summarized in Table 1 and some DSC curves areexemplified in Fig. 3. All three compounds (P26, P50, andP62) showed melting transitions during the first heatingscan (see Table 1). In the second heating scans, congenersP50 and P62 showed only glass transitions at −13.9 and20.1°C, respectively. For congener P26, somewhat differentbehavior was observed, as more complex crystallization-melting-crystallization (cmc) sequences were seen prior to

Table 3 Environmentally relevant properties of polychlorobornanes

A. Compound codesa,b, CAS registry numbers and constant propertiesComp CASNRc WM Mp Cd TM K ΔSf

d Vbe DBf KaccWf,g

P 26 142534-71-2 413.79 154.1 427.3 57.66 286.4 15.92 25P 50 66860-80-8 448.24 132.9 406.1 51.73 298.2 15.90 26.5P 62 154159-06-5 448.24 148.6 421.8 51.25 298.2 16.63 26.5B. Temperature correction coefficients for PL, SW, KOW, and H (Eq. (1))

VPLEST WATSOLU TDLKOW Conversionh

Comp Apl Bpl As Bs Aow Bow Ah BhP 26 10.16 3791 −0.347 1448.3 6.099 145.48 10.52 364.4P 50 10.20 3952 −0.591 1230.9 5.981 130.17 10.72 2404P 62 10.21 4001 −0.616 1223.7 5.981 93.70 10.83 2778C. Estimation of degradation rates (k(i)) at 25°C from energy dataComp ΔE0 (kJ mol−1)i ΔHfor (kJ mol−1)j Biostabilityk Photostabilityl Degradation rates h−1

k(p) k(h) k(b)P26 −230.64 15.95 214.69 28.8 0.000058 0.000050 0.000250P50 −245.68 0.00 245.68 9.40 0.000173 0.000040 0.000250P62 −187.68 38.52 149.16 1.2m 0.001355 0.000100 0.000400D. Degradation rates (R(i)) and reference half-life times (HL(i)) in four compartments at 25°CDegradation rates in media from k(i) valuesg Degradation (reaction) half-life time valuesComp R(1) h−1 R(2) h−1 R(3) h−1 R(4) h−1 HL(1) h HL(2) h HL(3) h HL(4) hP26 5.78E−05 7.56E−05 2.54E−04 2.50E−05 12,000 9,161 2,727 27,720P50 1.73E−04 7.53E−05 2.53E−04 2.42E−05 4,000 9,207 2,736 28,676P62 1.36E−03 2.46E−04 4.08E−04 4.17E−05 300 1,700 1,000 10,000

a Andrews and Vetter (1995)b Parlar et al. (1995)c Vetter and Oehme (2000)d Values from DSCe Ruelle et al. (1991)f Ruelle (2000)g Paasivirta (2006) (Increment value for Cl as associating group)h (Ah=Apl−As and Bh=Bpl−Bs)i Vetter and Scherer (1998)j Heimstad et al. (2001)k (ΔHfor−ΔE0 )l Coelhan and Maurer (2005)mDefault value

198 Environ Sci Pollut Res (2009) 16:191–205

melting transition at 150°C. This cmc sequence was nicelyrepeatable if additional heating–cooling cycles were madeon the same sample. The measurements, in which thetoxaphene samples were heated to 270°C, the decomposi-tion initiated typically above 230°C for P26 and P50, and ata slightly lower temperature (205°C) for P62, as can beseen in Fig. 3 and Table 1.

Modeling procedure The model runs with FATEMODchain, including the properties of environments (seeTable 2) and substances (see Table 3), and started usingdefault emissions (in rotation) of P26, P50, and P62 (inkg h−1) during the spread of insecticide (toxaphene mixture)to soil/plants at the two source areas (Table 4). Two sets of

Fig. 3 DSC scans of compounds P26 (1), P50 (2), and P62 (3)measured from 25°C to the decomposition temperature with a heatingrate of 10°C/min

Table 4 Modeling atmospheric transport of P26, P50, and P62 in June from sources 1 and 2 to the Barents Sea

Route of the major air trajectory P26 kg h−1 P50 kg h−1 P62 kg h−1

t °C FrFl Emission Advection Emission Advection Emission Advection

Case A (high emission)Source 1 20 0.4 40 (air) 10.1 (air) 176 (air)

400 (plants) 47.32 1010 (plants) 97.04 1760 (plants) 137.5Ratio factor from anal. 4 4 10.1 10.1 17.6 17.6NWB 18 0.2 18.93 18.44 38.81 37.81 67.64 42.39Source 2 30 0.3 40 (air) 10.1 (air) 176 (air)

400 (plants) 40.968 1,010 (plants) 97.6165 1,760 (plants) 136.5302KAZ 25 0.5 12.29 12.064 29.285 27.755 40.959 31.939CWRU 20 0.5 6.032 5.988 13.877 13.564 15.969 13.066NEVA 18 0.666 6.683 6.5 14.345 13.285 15.011 11.605CFI 15 0.77 4.329 4.168 8.847 8.005 7.729 5.898KemR 8 0.834 3.209 2.996 6.164 5.277 4.541 3.342F/L 7 0.909 2.498 2.33 4.401 3.769 2.787 2.054

Adv/water→ 6.40E−04 1.12E−03 0.001Barents Sea 5 Emis/air→ 2.118 3.426 1.867

Emis/Water→ 5.82E−04 1.02E−03 6.62E−04Case B (low emission)Source 1 20 0.4 19 (air) 27 (air) 19 (air)

190 (plants) 22.483 270 (plants) 25.942 190 (plants) 14.839Ratio factor from anal. 1.9 1.9 2.7 2.7 1.9 1.9NWB 18 0.2 8.993 8.761 10.377 10.109 5.936 4.577Source 2 30 0.3 19 (air) 27 (air) 19 (air)

190 (plants) 19.46 270 (plants) 26.096 190 (plants) 14.733KAZ 25 0.5 5.838 5.73 7.829 7.42 4.422 3.448CWRU 20 0.5 2.865 2.844 3.71 3.626 1.724 1.411NEVA 18 0.666 3.174 3.088 3.835 3.551 1.621 1.253CFI 15 0.77 2.056 1.98 2.365 2.14 0.834 0.637KemR 8 0.834 1.524 1.423 1.648 1.411 0.49 0.361F/L 7 0.909 1.187 1.107 1.177 1.008 0.301 0.222

Adv/water→ 3.04E−04 3.00E−04 7.87E−05Barents Sea 5 Emis/air→ 1.006 0.916 0.202

Emis/water→ 2.72E−04 2.76E−04 7.15E−05

The emissions were estimated from two possible use cases: A (high estimate) and B (low estimate). Emissions are to air and at sources mainly toplants; advections are to air, from Finnmarken/Lappland to Barents Sea also with water. Modelled advection from a model box is multiplied withestimated fraction of flow (FrFl) to the next box and then taken as emission there. Trajectories join together at the NEVA area, where the twoflows were summarized

Environ Sci Pollut Res (2009) 16:191–205 199

emissions of the compounds (A and B) were chosen basedon our analysis results. Set A (high level) was chosen fromanalysis of Western toxaphene product camphechlor(Nikiforov et al. 1999). Only a preliminary analysis reporton polychlorobornanes in Soviet polychloropinene formu-lation was published (Nikiforov et al. 2004), and no otherresults were found in literature. Therefore, to explore set B(low level), five samples of Russian chloroterpene insecti-cide mixtures (from two polychlorocamphenes and threepolychloropinenes) were analyzed in present work. Con-sidering the toxaphene production numbers, the meanconcentrations of compounds in formulations were usedas ratio factors between A and B sets in a modelingschedule (see Table 4). In application, the spraying wasassumed to cause 100 parts of chlorobornane to plants andten parts instantly evaporate to air in all cases. These

default numbers were multiplied by ratio factors to givemodel case emission in kilograms per hour.

Mean concentrations — ratiofactors

P26(% w/w)

P50(% w/w)

P62(% w/w)

A. High level set (Westerncamphechlor)

0.40 1.01 1.76

B. Low level set (Russianchloroterpenes)

0.19 0.27 0.19

The advection with air from source area was thenmultiplied by fraction emitted by air flow to next FATEMOD chain box and the emission in that box. Then, Level3 estimation gave the steady state advection with air to thethird model box, etc. The modeled emissions from NWB

Table 5 Summary of model results for the Barents Sea: FATEMOD level 3 calculation, case A, emissions to air and water included

Fate of Parlar 26 at Barents Sea (at 5°C): case A, high emissionSteady state emissions and concentrationsEmiss. to air 2.1180 kg h−1

Emiss. to water 5.82E−04 kg h−1

AIR ng L−1 WATER ng L−1 SOIL ng g−1 dw SEDIM. ng g−1 dw FISH ng g−1 ww2.15E+00 4.02E−03 6.94E−03 4.08E−03 7.26E−02Bioaccumulation—only free (truly dissolved) substance includeda

Plankton Zoobenthos Planktovores PiscivoresBAF BIOlw/Cw 3,489,538 3,152,711 3,708,286 6,032,566Log(10)BAF 6.543 6.499 6.569 6.781Cwater ng L−1 ng g−1 lw ng g−1 lw ng g−1 lw ng g−1 lw4.02E−03 14.02 12.67 14.90 24.24Fate of Parlar 50 at Barents Sea (at 5°C): case A, high emissionSteady state emissions and concentrationsEmiss. to air 3.4260 kg h−1

Emiss. to water 1.02E−03 kg h−1

AIR ng L−1 WATER ng L−1 SOIL ng g−1 dw SEDIM. ng g−1 dw FISH ng g−1 ww3.31E−05 3.08E−02 3.18E−02 2.73E−02 4.48E−01Bioaccumulation—only free (truly dissolved) substance includeda

Plankton Zoobenthos Planktovores PiscivoresBAF BIOlw/Cw 2,980,618 2,656,188 3,070,538 4,958,785Log(10)BAF 6.474 6.424 6.487 6.695Cwater ng L−1 ng g−1 lw ng g−1 lw ng g−1 lw ng g−1 lw3.31E−05 92.72 81.73 94.48 152.59Fate of Parlar 62 at Barents Sea (at 5°C): case A, high emissionSteady state emissions and concentrationsEmiss. to air 1.8670 kg h−1

Emiss. to water 6.02E−03 kg h−1

AIR ng L−1 WATER ng L−1 SOIL ng g−1 dw SEDIM. ng g−1 dw FISH ng g−1 ww1.44E−05 8.98E−03 5.94E−03 6.58E−03 1.89E−01Bioaccumulation—only free (truly dissolved) substance includeda

Plankton Zoobenthos Planktovores PiscivoresBAF BIOlw/Cw 4,127,836 3,787,636 4,539,723 7,454,399Log(10)BAF 6.616 6.578 6.657 6.872Cwater ng L−1 ng g−1 lw ng g−1 lw ng g−1 lw ng g−1 lw8.98E−03 37.09 34.03 40.79 66.97

a Voutsas et al. (2002)

200 Environ Sci Pollut Res (2009) 16:191–205

and CWRU areas were added to obtain the model emissionfor the Neva area to continue modeling emission chaintowards the Barents Sea. Finally, modeled Level 3advections from Finnmarken/Lappland with air and waterwere converted to emissions into Barents Sea air and waterfor modeling the steady state there.

Validation of the model results At end of the FATEMODchain, Level 3 modeling resulting in water columns of theBarents Sea box produced concentrations in water forbioaccumulation evaluation. Statistically developed BAFequations for highly hydrophobic POPs (chlorobornanes),using their KOW’s as a factor (accuracy order of magnitudeor better for truly dissolved compound in water; Voutsas etal. 2002). In this way, Level 3 modeled estimates of lipid-based concentrations of the substances in four trophic levels

of aquatic species were obtained (Tables 5, 6, and 7). Themodel estimates were compared with analytical results fromaquatic biota sampled in 1993 (Oehme et al. 1995; Karlssonet al. 1997) and in August 1997 (Wolkers et al. 2000;Foreid et al. 2000; Soermo et al. 2006). Start of modelemission at source areas was in early summer 1997,transport in June–July. The modeled and analyzed levelsin Barents Sea species are collected in Table 7. The levelswere all in ng g−1 lw class. The model results (ng g−1 lw):4–95 in lipid of planktovores and 7–150 in lipid ofpiscivores, were in fair agreement with the analysis resultsfrom August 1997: 21−31 in Themisto libellula (chatka),26–42 in Boreocadus saida (polar cod), and 5–27 in Gadusmorhua (cod) liver. From the six values modeled for Highemission, two (P26) were below and the other four (P50,P62) were above the mean analyzed levels in chatka and

Table 6 Summary of the model result for the Barents Sea: FATEMOD level 3 calculation, case B, emissions to air and water included

Fate of Parlar 26 at Barents Sea (at 5°C): case B, low emissionSteady state emissions and concentrationsEmiss. to air 1.0060 kg h−1

Emiss. to water 2.79E−04 kg h−1

AIR ng L−1 WATER ng L−1 SOIL ng g−1 dw SEDIM. ng g−1 dw FISH ng g−1 ww1.02E−05 1.91E−03 4.58E−03 1.94E−03 3.45E−02Bioaccumulation—only free (truly dissolved) substance includeda

Plankton Zoobenthos Planktovores PiscivoresBAF BIOlw/Cw 3,489,538 3,152,711 3,708,286 6,032,566Log(10)BAF 6.543 6.499 6.569 6.781Cwater ng L−1 ng g−1 lw ng g−1 lw ng g−1 lw ng g−1 lw1.91E−03 6.66 6.02 7.08 11.51Fate of Parlar 50 at Barents Sea (at 5°C): case B, low emissionSteady state emissions and concentrationsEmiss. to air 0.9160 kg h−1

Emiss. to water 2.72E−04 kg h−1

AIR ng L−1 WATER ng L−1 SOIL ng g−1 dw SEDIM. ng g−1 dw FISH ng g−1 ww8.84E−06 8.23E−04 8.50E−03 7.30E−03 1.29E−01Bioaccumulation—only free (truly dissolved) substance includeda

Plankton Zoobenthos Planktovores PiscivoresBAF BIOlw/Cw 2,980,618 2,656,188 3,070,538 4,958,785Log(10)BAF 6.474 6.424 6.487 6.695Cwater ng L−1 ng g−1 lw ng g−1 lw ng g−1 lw ng g−1 lw8.23E−04 24.52 21.85 25.26 40.80Fate of Parlar 62 at Barents Sea (at 5°C): case B, low emissionSteady state emissions and concentrationsEmiss. to air 0.2020 kg h−1

Emiss. to water 2.72E−05 kg h−1

AIR ng L−1 WATER ng L−1 SOIL ng g−1 dw SEDIM. ng g−1 dw FISH ng g−1 ww1.56E−06 9.72E−04 6.43E−04 7.12E−04 2.04E−02Bioaccumulation—only free (truly dissolved) substance includeda

Plankton Zoobenthos Planktovores PiscivoresBAF BIOlw/Cw 4,127,836 3,787,636 4,539,723 7,454,399Log(10)BAF 6.616 6.578 6.657 6.872Cwater ng L−1 ng g−1 lw ng g−1 lw ng g−1 lw ng g−1 lw9.72E−04 4.01 3.68 4.41 7.25

a Voutsas et al. (2002)

Environ Sci Pollut Res (2009) 16:191–205 201

polar cod. Values modeled for Low emission assumptionwere all lower than the analyzed ones (see Table 7).Roughly, levels of substances in the Barents Sea biota 1997were in agreement between modeled and analyzed ones.Any validation about overall camphechlor/polychloropi-nene ratio, however, was not achieved. Obviously, we hadtoo few data for firm validation. However, feasibility of thechained single-box model for long-range atmospherictransport to chlorobornane’s case was obvious. ModeledP26 levels were below the observed ones in both High andLow emission cases, non-similar to cases for P50 and P62(in High case below, in Low case above). This indicatesthat the value of photolysis rate of P26 in air had beenoverestimated (although predicted unusually low by us).

4 Discussion

The aim of this study as the first trial of this chain modewas to elucidate the model but not to validate the procedurecompletely. The modeling results indicate that the applica-tion of simple multimedia catchment boxes chained on apredicted trajectory is a useful method for estimation ofvolatile airborne persistent chemical exposure to biota inremote areas. For hazard assessment of these pollutants,their properties, especially temperature dependences, mustbe estimated by a reasonable accuracy, which can beachieved by using measurements in laboratory with puremodel compounds and estimation of properties by thermo-dynamic QSPR methods. The property parameters can be

validated by comparing their values at environmentaltemperature ranges using measured or QSPR-estimatedvalues derived by independent methods. Using chainedboxes can be more suitable than the previously used globalparallel zones method, provided that the main airflowtrajectories and properties of transported pollutants arepredictable enough. Inconsistency between modeled andobserved (literature) values appeared only in P26 levels,which were below the observed ones also found by Highemission assumption. This indicates some error in param-eter determinations. Maybe the very slow photodegradationrate of P26 in atmosphere was still evaluated as being toofast.

5 Conclusions

Long-range air transport modeling of persistent, especiallyphoto-resistant organic compounds using a chain of jointsimple boxes of catchment environments is a feasiblemethod to predict concentrations of pollutants at the targetarea. This is justified from model results compared withanalytical measurements in Barents Sea biota in August1997: two from three resulting sets, those for P50 and P62,were consistent with assumed high and low emission sets.For set P26, both emission defaults resulted below theobserved values. This inconsistency might be caused byoverestimation of the photolysis rate value, although ifevaluated by us lower it would be unusually small.Anyhow, the order of magnitude level was similar in

Table 7 Concentrations of toxaphene congeners P26, P50, and P62 to aquatic species from the Barents Sea

Observed values (analysis results from literature)Location Barents Sea East Svalbard JarfjordSampling year 1993 1997 1997 1997Ref. a, b c c, d eSpecies Cod Themisto Polar cod CodTrophic level Piscivore Planktovores Piscivores PiscivoresOrgan Liver Whole chatka Whole fish LiverExtr. lipid % 52.2 9.01 12.1 60.5P26 ng g−1 lw 44.1 21.1 25.8 22P50 ng g−1 lw 126.4 31.1 41.6 26.5P62 ng g−1 lw 52.9 23.3 27.5 4.5Modeled values by FATEMOD chain for summer 1997

A. High emission B. Low emissionPlanktovores Piscivores Planktovores Piscivores

P26 ng g−1 lw 14.9 24.2 7.1 11.5P50 ng g−1 lw 94.5 152.6 25.3 40.8P62 ng g−1 lw 40.8 67.0 4.4 7.3

a Oehme et al. (1995)b Karlsson et al. (1997)cWolkers et al. (2000)d Soermo et al. (2006)e Foreid et al. (2000)

202 Environ Sci Pollut Res (2009) 16:191–205

modeled (planktovore and piscivore) and observed (The-misto and polar cod) values in 1997 lipid samples (seeTable 7). The obtained result data were too limited toprovide firm validation, but sufficient to justify practicabil-ity of the method, which prompts us to make more studiesbased on this modeling system.

6 Recommendations and perspectives

For assessment of the risk of environmental damages,chemical fate determination is an essential tool for chemicalcontrol, e.g., for EU following the REACH rules. Thepresent conclusion of applicability of the chained single-box multimedia modeling can be validated by furtherstudies using analyses of emissions and target biota invarious other cases. To achieve a useful methodology inthis type of study, fate models built with the automatic stepsof databases are essential for most calculations and outputs,and are accessible for all professional chemical controls.Our FATEMOD program catchments in environments andcompound properties listed in a database is a feasible toolfor local, regional, and, according to our present test results,for global exposure predictions. As an extended use of thismodel, emission estimates can be achieved by reversedmodeling from analysis results of samples corresponding tothe target area. Chained single-box long-range transportmodels can also work well for persistent hydrophiliccompounds in watercourses.

References

AMAP (2002) Persistent organic pollutants in the Arctic. Monitoringand Assessment Programme. Oslo, 309 pp

Andrews P, Vetter W (1995) A systematic nomenclature system fortoxaphene congeners Part 1: chlorinated bornanes. Chemosphere31(8):3879–3886

Barrie L, Bidleman T, Dougherty D, Fellin P, Grift N, Muir D,Rosenberg B, Stern G, Toom D (1993) Atmospheric toxaphene inthe high Arctic. Chemosphere 27(10):2037–2046

Bidleman TF, Leone AD (2004) Soil–air exchange of organochlorinepesticides in the Southern United States. Environ Pollut 128:49–57

Bidleman TF, Olney CE (1975) Long range transport of toxapheneinsecticide in the atmosphere of the west North Atlantic. Nature(London UK) 257(5526):475–477

Bidleman TF, Wideqvist U, Jansson B, Söderlund R (1987)Organochlorine pesticides and polychlorinated biphenyls in theatmosphere of southern Sweden. Atmos Environ 21:641–654

Bidleman TF, Leone AD, Falconer RI (2003) Vapor pressures andenthalpies of vaporization for toxaphene congeners. J Chem EngNews 48:1122–1127

Burhenne J, Haintzl D, Xu L, Vieth B, Alder L, Parlar H (1993)Preparation and structure of high-chlorinated bornane derivativesfor the quantification of toxaphene residues in environmentalsamples. Fresenius J Anal Chem 346:779–785

Chandurkar PS, Matsumura F (1979) Metabolism of toxicant B andtoxicant C of toxaphene in rats. Bull Envir Contamin Toxicol 21(1):539–547

Chandurkar PS, Matsumura F, Ikeda T (1978) Identification andtoxicity of toxicant AC, a toxic component of toxaphene.Chemosphere 7:123–130

Coelhan M, Maurer M (2005) Synthesis of two major toxaphenecomponents and their photostabilities. J Agric Food Chem58:10105–10112

El-Sebae AH, Abou Seid M, Saleh MA (1993) Status and environ-mental impact of toxaphene in the Third World—a case study ofAfrican agriculture. Chemosphere 27(10):2063–2072

Fedorov LA (1999) Persistent organic chemicals in the former SovietUnion. Environ Pollut 105:283–287

Foreid S, Rundberget T, Severinsen T, Wiig Ö, Skaare JU (2000)Determination of toxaphenes in fish and marine mammals.Chemosphere 41:521–528

Frenzen G, Hainzl D, Burhenne J, Parlar H (1994) Structureselucidation on the three most important toxaphene congeners byx-ray analyses. Chemosphere 28:2067–2074

Grain CF (1990) Tables 14.4 and 14.5, Eqn 14–25 (Chapter 14): vaporpressure. In: Lyman WJ, Reehl WF, Rosenblatt DH (eds)Handbook of chemical property estimation methods. ACS,Washington, DC

Harner T, Bidleman TF, Jantunen LMM, Mackay D (2001) Soil–airexchange model of persistent pesticides in the United Statescotton belt. Environ Toxicol Chem 20(7):1612–1621

Heimstad ES, Smalås AO, Kallenborn R (2001) Environmental fate ofchlorinated bornanes estimated by theoretical descriptors. Che-mosphere 4:665–674

Jantunen LM, Bidleman TF (2003) Air–water gas exchange of toxa-phene in Lake Superior. Envir Toxicol Chem 22(6):1229–1237

Jantunen LMM, Bidleman TF, Harner T, Parkhurst WJ (2000)Toxaphene, chlordane and other organochlorine pesticides inAlabama air. Environ Sci Technol 34:5097–5015

Kallenborn R, Oehme M, Vetter W, Parlar H (1994) Enantiomerselective separation of toxaphene congeners isolated from sealblubber and obtained by synthesis. Chemosphere 28(1):89–98

KarlssonH, OehmeM,Klungsöyr J (1997) 4,5-Dichloro-chlordene, a newsynthetic internal standard for the quantification of toxaphene andchlordane congeners in fish from the Barents Sea and North Atlantic.Application and experience. Chemosphere 34(5–7):951–963

Lahtinen M, Paasivirta J, Nikiforov VA (2006) Evaluation of entropiesof fusion of polychlorinated naphthalenes by model congeners: aDSC study. Thermochimica Acta 447:5–12

Lammel G, Klöppfer W, Semeena WS, Schmidt E, Leip A (2007)Multicompartmental fate of persistent substances. Comparison ofpredictions from multimedia box-models and a multicompart-ment chemistry-atmospheric transport model. Environ Sci PollutRes 14(3):153–165

Li YF, Bidleman TF, Barrie LA (2001) Toxaphene in the United States2. Emissions and residues. J Geophys Res [Atmospheres] 106(D16):17929–17938

Mackay D (1991) Multimedia Environmental Models. The FugacityApproach. L242, Lewis, Chelsea, MI

MacLeod M, Woodfine D, Brimacombe J, Toose L, Mackay D (2002)A dynamic mass budget for Toxaphene in North America. EnvirToxicol Chem 21(8):1628–1637

Matsumura F, Howard RH, Nelson JO (1975) Structure of the toxicfraction A of toxaphene. Chemosphere 5:271–276

McConnell LL, Kucklick JR, Bidleman TF, Walla MD (1993) Long-range atmospheric transport of toxaphene to Lake Baikal.Chemosphere 27(10):2027–2036

Murphy TJ, Mullin MD, Meyer JA (1987) Equilibration of poly-chlorinated biphenyls and toxaphene with air and water. EnvironSci Technol 21:155–162

Environ Sci Pollut Res (2009) 16:191–205 203

Nicholson HP (1968) Water contamination by pesticides. J Southeast-ern Section, Am Water Works Assoc 32(1):8–20

Nicholson HP, Grzenda AR, Teasley JI (1968) Water pollution byinsecticides. J Southeastern Section, Am Water Works Assoc 32(1):21–27

Nikiforov VA, Karavan VA, Miltsov SA (1999) Analysis ofcamphechlor. Organohalogen Compounds 41:601–604

Nikiforov VA, Karavan VA, Miltsov SA (2000a) Relative retentiontimes of chlorinated monoterpenes. Chemosphere 41(4):467–472

Nikiforov VA, Karavan VA, Miltsov SA (2000b) Relative retentiontimes of 122 polychloroterpenes. Organohalogen Compounds50:268–271

Nikiforov V, Trukhin A, Kruchkov F, Kiprianova A, Miltsov S,Kallenborg R (2004) Isolation of pure enantiomers of toxaphenecongeners via hydrochlorination and chlorination of pinene andcomposition of Soviet polychloropinene. Organohalogen Com-pounds 66:456–461

Norstrom RJ, Simon M, Muir DCG, Schweinburg RE (1988)Organochlorine contaminants in Arctic marine food chains:identification, geographical distribution and temporal trends inPolar bears. Environ Sci Technol 22:1063–1071

Oehme M, Svendsen JS, Roland K, Klungsöyr J (1995) Character-ization and application of synthetic 4,5-dichlorochlordene asinternal standard for the quantification of toxaphene andchlordane congeners in fish from the Arctic. OrganohalogenCompounds 26:333–338

Paasivirta J (2006) Estimation of vapour pressure, solubility in water,Henry’s law function, and Log Kow as a function of temperaturefor prediction of environmental fate of chemicals. In: KungolosA, Brebbia AA, Samaras CP, Popov V (eds) Environmentaltoxicology. WIT transactions on biomedicine and health. vol. 10.WIT, Southampton, pp 11–20

Paasivirta J, Rantio T (1991) Chloroterpenes and other organo-chlorines in Baltic, Finnish and Arctic wildlife. Chemosphere22(1–2):47–55

Paasivirta J, Rantio T, Koistinen J, Vuorinen PJM (1993) Studies ontoxaphene in environment. II. PCCs in Baltic and Arctic Sea andlake fish. Chemosphere 27(10):2011–2015

Paasivirta J, Sinkkonen S, Mikkelson P, Rantio T, Wania F (1999)Estimation of vapour pressures, solubilities and Henry’s lawconstants of selected persistent organic pollutants as functions oftemperature. Chemosphere 39(5):811–832

Paasivirta J, Sinkkonen S, Rantalainen A-L, Broman D, Zebühr Y(2002) Temperature dependent properties of environmentallyimportant synthetic musks. Environ Sci Pollut Res 9(6):345–356

Pacyna JM, Brorström-Lunden E, Runge E, Paasivirta J, Münch J,Fudala J, Calamari D, Broman D (1999) Environmental cyclingof selected persistent organic pollutants (POPs) in the BalticRegion. ENV4-CT96-0214/EU. Executive Final Summary Re-port, Annex 1: Technical Report and CD-ROM including: 1)Technical Report, 2) The Emission Database, 3) The PopcyclingModel and 4) The Environmental Database. NILU, Kjeller

Parlar H, Angerhöfer D, Coelhan M, Kimmel L (1995) HRGC andHRGC-ECNI determination of toxaphene residues in fish with a new22 components standard. Organohalogen Compounds 26:357–362

Pyysalo H, Antervo K (1985) GC profiles of chlorinated terpenes(toxaphenes) in some Finnish environmental samples. Chemo-sphere 14(11–12):1723–1728

Rantio T, Paasivirta J, Lahtiperä M (1993) Studies on toxaphene inenvironment. I. Experiences on analytical behavior of PCCs. Studiesincluding pulp mill wastes. Chemosphere 27(10):2003–2010

Ruelle P (2000) The n-octanol and n-hexane/water partition coeffi-cient of environmentally relevant chemicals predicted from the

mobile order and disorder (MOD) thermodynamics. Chemo-sphere 40:457–512

Ruelle P, Kesselring UW (1997) Aqueous solubility prediction ofenvironmentally important chemicals from the mobile orderthermodynamics. Chemosphere 34:275–298

Ruelle P, Rey-Mermeth C, Buchmann M, Ho N-T, Kesselring UW,Huyskens PL (1991) A new predictive equation for the solubilityof drugs based on the thermodynamics of mobile disorder. PharmRes 8:840–850

Saleh MA (1991) Toxaphene: chemistry, biochemistry, toxicity andenvironmental fate. Rev Environ Contam Toxicol 118:1–85

Schwarzenbach RP, Cscwend PM, Imboden DM (1993) Environmen-tal organic chemistry. Wiley, New York, pp 80–81

Semeena WS, Lammel G (2003) Effect of various scenarios uponentry of DDT and γ-HCH into the global environment on theirfate as predicted by a multicompartment chemistry-transportmodel. Fresenius Environ Bull 12:925–939

Sinkkonen S, Paasivirta J (2000) Degradation half-life times ofPCDDs, PCDFs and PCBs for environmental fate modelling.Chemosphere 40:943–949

Sleeswijk AW (2006) GLOBOX—a spatially differentiated multime-dia fate and exposure model. Incl. instructions for download ofGLOBACK. Environ Sci Pollut Res 13(2):143–144

Soermo EG, Pettersvik Salmer M, Jenssen BM, Hop H, Bek K,Kovacs KM, Lydersen C, Falk-Petersen S, Gabrielsen GW, LieE, Skaare JU (2006) Biomagnification of polybrominateddiphenyl ether and hexachloromonocyclododecane flame retard-ants in the Polar Bear food chain in Svalbard, Norway. EnvirToxicol Chem 25(9):2502–2511

Stern GA, Muir DCG, Ford CA, Grift NP, Dewailly E, Bidleman TF,Walla ND (1992) Isolation and identification of two majorrecalcitrant toxaphene congeners in aquatic biota. Environ SciTechnol 26:1838–1840

Sundström G (1981) Toxafen in Svensk Miljö—Tillförsel viaLuftnedfall (in Swedish), NORDFORSK miljövårdserien publi-kation 1981(1):331–336

Swackhamer DL, McConnell LL, Gregor DJ (1993) Workgroup report onenvironmental transport and fate. Chemosphere 27(10):1837–1840

Swackhamer DL, Schottler S, Pearson RF (1999) Air–water exchangeand mass balance of toxaphene in the Great Lakes. Environ SciTechnol 33:3864–3872

Swanson KL, Hope H, Landrum PF (1978) A toxaphene component:2,5,6-exo-,8,8,9,10-heptachloro-dihydrocamphene. Acta CrystB34:3411–3414

Tabor EG (1966) Contamination of urban air through the use ofinsecticides. Trans NY Acad Sci 28(5):569–578

Tarhanen J, Koistinen J, Paasivirta J, Vuorinen PJ, Koivusaari J,Nuuja I, Kannan N, Tatsukawa R (1989) Toxic significance ofplanar aromatic compounds in Baltic ecosystem—new studies ofextremely toxic coplanar PCBs. Chemosphere 18(1–6):1069–1077

Trapp S,MatthiesM (1998) Chemodynamics and environmental modeling.An introduction. CemoS user’s manual and diskette. Springer, Berlin

Tribulovich VG, Nikiforov VA, Karavan SA, Bolshakov S (1994)Synthesis and characterization of toxaphene congeners. Organo-halogen Compounds 19:97–101

Turner WV, Engel JL, Casida JE (1977) Toxaphene components andrelated compounds: preparation and toxicity of some hepta-, octa-,and nonachlorobornanes, hexa- and heptachlorobornenes, and ahexachlorobornadiene. J Agric Food Chem 25(6):1394–1401

Vaz R, Blomqvist G (1985) Traces of toxaphene components inSwedish breast milk analyzed by capillary GC using ECD,electron impact and negative ion chemical ionization MS.Chemosphere 14(2):223–231

204 Environ Sci Pollut Res (2009) 16:191–205

Vetter W, Scherer G (1998) Molecular modelling: an interesting tool toexplain persistence and lability of compounds of technicaltoxaphene. Organohalogen Compounds 35:235–238

Vetter W, Oehme M (2000) Toxaphene. Analysis and environmentalfate of congeners. In: Paasivirta J (ed) The handbook ofenvironmental chemistry. Vol 3 Part K New types of persistenthalogenated compounds. Springer, Berlin, pp 237–287

Voldner EC, Li YF (1993) Global usage of toxaphene. Chemosphere27(10):2073–2078

Voldner EC, Li YF (1995) Global usage of selected persistent organicchemicals. Sci Total Environ 160/161:201–210

Voutsas E, Magoulas K, Tassios D (2002) Prediction of thebioaccumulation of persistent organic pollutants in aquatic foodwebs. Chemosphere 48:645–651

Wania F, Mackay D (1993a) Global fractionation and cold condensa-tion of low volatility organochlorine compounds in Polar regions.Ambio 22:10–18

Wania F, Mackay D (1993b) Modelling the global distribution ofToxaphene: a discussion of feasibility and desirability. Chemo-sphere 27(10):2079–2094

Wania F, Persson J, di Guardo A, McLahlan M (2000) ThePOPCYCLING-Baltic model—a non-steady state multicompart-ment mass balance model of the fate of persistent organicpollutants in the Baltic Sea environment. Norwegian Institute ofAir Research Report No. NILU OR 10/2000, Kjeller Norway

Wolkers H, Burkov IC, Lydersen C, Witkamp RF (2000) Chlorinatedpesticide concentrations, with an emphasis on polychlorinatedcamphenes (toxaphenes), in relation to cytochrome P450 enzymeactivities in harp seals (Phoca groenlandia) from the Barents Sea.Environ Toxicol Chem 19(6):1632–1637

Zell M, Ballschmiter K (1980) Baseline studies of global pollution II.Global occurrence of hexachlorobenzene and polychlorocam-phenes (Toxaphene*) (PCC) in biological samples. Fresenius ZAnal Chem 300:387–402

Environ Sci Pollut Res (2009) 16:191–205 205