Document KJKNR1eNqRw57OppJpZ40xvx0

........... M. rr IM14 RimKat IH tilvaf Hral.nn ! . - i ( , ; ^ [ ATMOSPHERIC TRANSPORT OF POLYCHL.OROBIPHENYLS TO THE NORTH ATLANTIC* CiiOKGi: R. Harvey and William G. Stkinhauhr Woods Hole Occunoprnptuc Institution, Woods Hole, Mnssiichusetis 02543. U.S.A. {First rt'ccinti 5 .iv.sr 1973 anti in fituti fnrui 22 November 19731 Abstract Concemru linns or polychlorobipheiiyls (PCBs) were measured in the atmosphere over the western North Atlantic. Iloth the particulate and vapor phase were extracted, usinp a new tut nr pier designed for this work. Previously PCI! levels rnnpcd from near the northeast coast or the U.S. to tl-U5 np nt" J on the Orand 13.inks. The PCB concentration in tire atmosphere decreases cspoieniiitliy with distance from land. It is concluded that wind transport is the major source olchlorinated hydrocarhmrs in the oecans. INTRODUCTION Polychlorobiphenyl (PCB) concentrations have been measured in the water and organisms of Ihc open Atlantic Ocean (Harvey et at., 1973). It has generally been assumed that PCBs would be delivered to the open ocean mainly by the wind (Nisbet and Sarofim. 1972) simi lar to DDT (Seba and Prospero, 1971) with fluvial and sewage inputs contributing smaller quantities (Corcoran and Lee. 1970; Holden. 1970). However, no atmospheric PCB measurements have been reported. Risebrough (1972) was unable to detect PCBs on equa torial Atlanlic dust collected on Barbados (Risebrough et at., 1968). Therefore, it seemed likely that PCBs were transported mainly in the vapor or aerosol phase and we set out - to design a sampling system capable of trapping both particulate and vapor phase chlor inated hydrocarbons. Atmospheric PCB measurements were important to our studies because we had previously observed no inshore-open ocean gradient of PCBs in the sur face waters of the North Atlantic (Harvey et al., 1973). Thus, we wished to determine if the atmospheric concentrations of PCB reflected this observation. In this paper we present a new atmospheric sampling technique, and the results of our PCB measurements at four stations over the western North Atlantic. The sampling was performed during the first half of 1973. . PROCEDURES. The sampling system consisted of a Hurricane Air Pump {Bendix Corp.)fitled with a 102 mm filter holder on the intake port. The exhaust pipe was fitted with several meters of extension so as not to interfere with the intake. A 0.3 jtm glass fiber filter, which had been baked at 500 for 12 h, was used on the intake. The filter holder chamber, behind the filter was filled with 1/4 in. ceramic saddles (distillation column packing, Fisher Scientific) which were coated with 0.25 per cent OV-17 silicone oil. The pump was run on low speed and generally achieved a flow rate of 22-24 ft3 min" *. The saddles had to be cleaned initially by distilling several liters of water through a column packed with about 500 g of the commercial product. The OV-17 was purified by applying 1,25 g to the top of a 1.5 x I5cm column of silica. Interfering impurities were * Woods Hale Oceanographic Institution Contribution No. 3173. mpwmjK uji Diijiinaw^s^nfmi3ji.niRwiwiKw.iujiiyifaw|p>Uiiii at* 0500289 STLCOPCB4044459 ?7ti CroRCr R Hahvi v and Will JAM G- Sn.lNllAurx eluted with SO cm1 of hexane. The OV-J7 was chi ted with 150cms of benzene. The benzene solution was diluted with 100 cm* of hexane and 500 g of steam cleaned saddles were added. The solvent was evaporated to dryness on a rotary evaporatory at 50" with a water aspiratory. The coated saddles could be stored in glass jars or used immediately. About 125garc required to fill the filler holder chamber. Recovery of the chlorinated hydrocarbons from the coated saddles was accomplished by shading with two I Of) era'1 portions of hexane in a separatory funnel. The hexane solu tion was dried with sodium sulfate and concentrated to 10 cm* by distillation. The extract was applied to the top 1.5 x 15cm column of silica and the chlorinated hydrocarbons were eluted with 80 cm3 of hexane. The hexane cluatc was concentrated to I cm3 in a Kuderna-Onnish concentrator. If necessary, the hexane was further concentrated with a stream of Floristl-purgcd nitrogen. No further cleanup was necessary before gas chromat ography, The concentrated extract was analyzed on a gas chromatograph equipped with a *3Ni electron capture detector. The column was a 2m x 2mm (i.d.) glass coil packed with a 1:1 mixture of 8 per cent QF-l and 2 per cent OV-17 on 100-120 mesh Anakrom. The oven temperature was 190". The chromatogram was quantified by comparing the heights of seven peaks in the sample with the same matching peaks in a standard of commercial Aroclor 1254 (Monsanto) as shown in Fig. 1. The blank (a) was produced by leaving the entire apparatus at the sampling site for 24 h without running the pump. The saddles were then extracted and chromatographed as described above. The sample chromatogram (fy) I l Fi;. 1. Typical PCB chromatogram of a Orand Banks air sample <b). The standard Aroclor 1254 fo shown in (c) and the blank is shown in (&}. The teen dotted peaks were used Xor the quantification. rpnuaifcj 05o*9o STLCOPCB4044460 AlaiiOKphcric transport of polyehlorolripbcnyla 779 is from the Grand Banks.iTlic presence of lhc PCBs was confirmed by pcrchlorinalion to decnchlorobiphenyl according to the procedure of Armour (1973). The dccachlorobi- phcnyl synthesized from the Bermuda siimpics was further confirmed by mass spectro metry. Extraction of the glass fiber filters with hexane followed by concentration arid chromat ography revealed PCI) concentrations which Tanged from undetectable in the Grand Banks samples to 0.04 ng m " in the Vineyard Sound samples. The efficiency of trapping was tested by applying known quantities of 3,4,2'-trichlorobi- phcnyl; 2,4.5,2.'.4',5'-hexachloTObiphcnyl (Antitabs, Inc.) or Aroelor 1254 (Monsanto Co.) on the glass fiber filter. The average recovery after 20 h of pumping was 70 per cent for the pure isomers or the mixture. The sampling stations arc shown in Fig- 2. Vineyard Sound (Massachusetts) samples were collected from lhc roof of the pilot cabin of the W.H.O.J. vessel R/V Astcrlas. The Georges Bank collections were made from the 01 deck of the R/V Albatross l V(National Marine Fisheries Service). The Bermuda samples were taken from a 20-m tower on the south-west shore of the island. Finally, the collections on the Grand Banks were performed from the exploratory drilling rig Sedneth 1 of Amoco Canada Ltd. at Station Razorbill. Procedural blanks from each station averaged 0.03 ngm-3 for 500 m3. RESULTS The analytical results for the four stations are set out in Table 1. In addition, the DDT family (r-DDT) was detected in the Bermuda samples. The concentration of DDT in the four samples decreased from 5.3 x 10~ '1 gm'J on 12 February, to 1.7 x 10-l,gm-3on 15 February (1973). The most likely source was the southeastern US. cotton fields directly to the west. At the other sampling sites the t-DDT concentration was less than I x 10" 13 gm"J. The PCB/r-DDT ratio was also found to be greater than 30 in the sur face water (Harvey ct pf,, 1973). The PCB concentrations were highest in Vineyard Sound which lies only 150-250 km from the nearest windward industrial sources, i.e. the BostonHartford-Ncw York-New Jersey complex. The concentrations of PCB over the Grand Banks which are over 2000 km from the industrial complex of the northeast United States were 100 times less than in Vineyard Sound. In fact, when the average concentrations at each station are plotted against the average distance ol the site from major industrial 7 BO Station Bcrroud;) (32 '2(TN: <vi 'dttrW) 132*2VN; 64 4tVW) (W20`N; M 40 W) George* l!;mk (4I"40*N; 67 JO'W) (4!*'4(rN: 6? 30 W) (41*40'N; 67'30`W) <4r40*N;6?\nYW) (41r*40N:6?'3CrW) Vtncyurd Sound (41-20`N; 70" SOW) Grand Banks (4516*N; 52"0H"W) (45^I6'N;52WW) (45C16'N; 52'08'W) (45I6'N; 52``0$'W) Gitmtir R. Hahvi y and Wu.i.iam O. Sii iniiahi h Tabic I. PCH conccntrmions over the xvcAlcrn North Atlantic VhXtK 11973* Sample volu me <m'> Wind direction 12 February 13 February 14 February 15 February 10 April I 3 April IS April 17 April 19 April 21 April 13 April 30 April 25 June 26 June 27 June 28 June 29 June 560 4X0 820 500 105 675 660 655 640 6S0 105 224 70 960 840 940 540 WNW W Variahte S NW NW NE NW sw sw sw sw ssw sw W5W wsw w PCH Tip m - (calc, a* Aroclor. 1 254) 0.5 0.4 0.16 0.15 1.4 0.R2 0.58 0.61 0.80 1.60 3.9 5.3 0.05 0.07 0.10 0.16 005 centers (Bermuda is equidistant from both northeast and southeast U.S. industrial com plexes) the seaward decrease is actually exponential (Fig. 3). Similar observations hav6 been made for wind transport of industrial particulates over the ocean (Folgcr, 1970) and1 bomb fallout over the United States after nuclear tests (Eisenbud and Harley. 1953). I WIWUI Fig. 3. Relationship of atmospheric PCB and distance from industrial sources. DISCUSSION In our opinion the collection method used can only approximate the relative con centrations of chlorinated hydrocarbons in the particulate and vapor-aerosol phase. Hyd rocarbons in the vapor or aerosol phase will easily pass through the glass fiber filter and be trapped in the silicone oil at an efficiency similar to the spiking experiments. Hydrocar bons adsorbed on solid particles which are trapped by the filter still have equilibrium \ I i V'WIHli STLCOPCB4044462 Atmojsplicrii: transport rtf pt'iychloropiphcnyfs 7X1 vapor pressures with tile surrounding air. At the high flow rales used IIrese iwlsorbcil hyd rocarbons could be entrained in the (low and swcpl into the vapor trap, The process is similar to that utilised in pas liquid chromatography. However, if we can assume that par ticulates are being collected throughout the sampling period, the observation that only I per cent or less of the loud chlorinated hydrocarbon concentration in the sampled air is (bund on the filler argues strongly against significant particulate transport. It seems likely that during the atmospheric lifetime of chlorinated hydrocarbons the vapor, nerosol and particulate states will equilibrate and exchange many times depending on temporal meteorological parameters. Thus, the slate in which they happen to be cap tured at n particular geographical location is relatively unimportant to considerations of total atmospheric transport. Thus, the concentrations reported in Table 1 should be regarded as total PCB m-J. The strong dilution and fallout gradient observed over the western North Atlantic is not reflected in the surface water concentrations. However, since the highest con centrations of suspended matter are in coastal waters {Manheim et a/.. 1970) adsorptive scavenging by organisms and sedimentation must deplete these waters of dissolved PCBs at rates comparable to the delivery. In contrast, only a small fraction of the discharged PCBs reach the open ocean, but that which does must Temain dissolved in the surface waters for a long time because oT the low concentration of suspended matter and the Slower sedimentation rate. The result of these processes is that the concentration of dis solved PCBs from near shore water and the open sea is about the same (Harvey et ai,, 1973). Sediments should reveal a gradient of PCBs from inshore to the abyssal plains. Analyses ofbottom feeders support this prediction (Harvey et a!-, 1974). . Since the relative unimportance ofriver transport of chlorinated hydrocarbons has been noted (Seba and Prospero, 1971; Hartung and Klinger. 1970; Corcoran and Lee, 1970; Seba and Corcoran, 1969) wind transport is the major source of chlorinated hydrocarbons in the oceans. Acknowledgements--We thank Drs. R, Duce and T. Bidleman, University of Rhode Island for permission end assistance in using their Bermuda tower. I>r. M. Grosslein made space available aboard R/V Albatross SK National Marine Fisheries Service. We are especially grateful to Mr. O. Medieros for collecting the samples on Athairms tv. Finally, we are indebted to Mr. Neil Greer (Amoco Canada Ltd.) for making accommodations available on the Sednctii l. This work was supported by the International Decade of Ocean Exploration. National Science Foundation, Gram GX-35212. REFERENCES Armour J. A. (1974) Quantitative perchlorination ofpolychlorinated biphenyls as a method for confirmatory resi due measurement and identification. J. Assoc. Off Agric. (in press). Corcoran E. F. and Lee C. C. (I97G) The distribution of chlorinated hydrocarbon pesticides in the Mississippi River Deha and adjacent Gulf of Mexico. Paper presented before American Dn'fniral Society^ 4 December. New Orleans, Louisiana. Eisenbud M. and Harley J. H. (1953) Radioactive dust from -nuclear detonations. Science 117, 141-147. Fotger D. W. {19701 Wind transport of land derived mineral, biogenic and industrial matter over the North Atlan tic, Peep Sea Res. 17. 337-352. * ,Hartung R. and Klinger G. W.< 1970) Concentration t DDT by sedimented polluting oils. Environ. ScL Tecknot. 4 407-410. Harvey G. R,, Slcinhaucr W, O. and Teal J. M. (1973) PolychlorobiphcnyU in North Atlantic Ocean water. Science 180,643--644. Harvey C. R.. Miklas H. P- Bowen V. T. and Slcinhaucr W. G. (1974) Observations on the distribution of chlor* inated hydrocarbons in Atlantic Ocean organisms. J. Mar. Res. (in press). Holden A. V. (1970) Source of polychlorinated biphenyl contamination in "the marine environment. Nature 228, 7B3-T84. * 782 Cji:orci: R. !Iarvj:v and William C. Sti;inu*ui-.k Manhi'im F. Meade R, H, and Bond Q. C\{IV70J Suspended nimicr in surface waters of tl*c Atluntic continen tal maipin front Crape Cod to the Florida Keys. Sricnrc 167, 37f-37f. Riwrhrouph H. W. Hitpgcit R. J., Griffin 3. J. ond Goldberg Si. >. (1968) Pesticides: Trans-Atlantic movements in the northeast trades. Srii-ncr 159, 1233-1235. Etisebrough R. W. {1972) Personal communication. Seba D. B. and Corcoran E. F. (1969) Surface slicks as concentrators of pesticides in (he marine environment. Pest. Monti. J. 3, 190-193. Seba P. p. and Prosper J. M. (1971) Pesticides in the lower atmosphere of the northern equatorial Atlantic Ocean. Atmospheric Bntlrontncni 5, 1043-1050. < * i i STLCOPCB4044464