Marine Pollution Bulletin, Volume 26, No.3, pp. 128-134, 1993.Printed in Great Britain.

Marine Pollution Bulletin

0025-326X/93 $6.00+0.00© 1993 Pergamon Press Ltd

Contamination of Coastal versus OpenOcean Surface WatersA brief meta-analysisW. JACKSON DAVISDepartment of Biology, University of California, Santa Cruz, C4. 95064; and The Environmental Studies Institute, *2309 Empire Grade, Santa Cruz, CA 95060, USA

The relative contamination of coastal and open oceansurface waters by several classes of marine pollutants iscompared by re-analysis of data published by GESAMP(1990) on the state of the oceans. This re-evaluation,together with analysis of recent IAEA data, suggests thatmarine pollutants can be divided into three broadcategories: 1. those that are more concentrated in theopen ocean environment than in coastal waters, includ-ing PCBs and (arguably) radioactive wastes; 2. thosethat are equally concentrated in open ocean and coastalwaters, including DDT and its metabolites; and 3. thosethat are somewhat more concentrated in coastal watersthan the open ocean, including chlorinated hydro-carbons and the heavy metals. The data do not supportthe conclusion drawn by GESAMP that althoughcoastal waters are contaminated, "the open sea is stillrelatively clean". Instead the data support the generalconclusion that, with respect to the contaminantsexamined here, open ocean waters are approximately ascontaminated as coastal waters.

There is widespread perception that coastal waters maysuffer point sources of significant pollution, but "theopen sea is still relatively clean" (GESAMP, 1990). Sucha perception reduces the motivation for global actionto preserve the marine environment. Moreover, this"dirty coasts clean seas" hypothesis is self-perpetuating,because it inhibits the research and monitoring of theopen ocean that could test the accuracy of the per-ception. The perception that the open ocean is stillrelatively clean therefore has both policy and scientificramifications.

The 'dirty coast clean seas' hypothesis originates inpart from a recent analysis of the health of the marineenvironment by GESAMP-the authoritative Group ofExperts on the Scientific Aspects of Marine Pollution(GESAMP), sponsored by several United Nations

*A California Non-Profit Corporation founded in 1981 for purposesof research and public education on scientific issues related to environ-mental policy.

128

agencies and issued by the United Nations Environ-mental Program (UNEP). This report notes that coastalregions are deteriorating seriously world-wide: "destruc-tion of beaches, coral reefs and wetlands, includingmangrove forests, as well as increasing erosion of theshore, are evident all over the world:' In contrast, theGESAMP report concluded, in summary, that "the opensea is still relatively clean".

The purpose of this brief communication is to presenta partial re-evaluation of the database on which theGESAMP summary conclusions are founded. This re-analysis does not support the 'dirty coast clean seas'hypothesis. Instead it suggests, for the several con-taminants examined, that the surface waters of theopen seas are approximately as contaminated as coastalwaters. This conclusion has implications for bothscientific and policy decisions in the area of marinepollution.

MethodsThe data on which the GESAMP conclusions are

based are contained in technical annexes to theGESAMP report (UNEP, 1990). The present analysisis confined to data on the concentration of marinecontaminants in coastal and open ocean surface waters,contained in the excellent review by Fowler (1990). Datapublished in the literature, and therefore in Fowler'sreview, are generally based on random or pseudo-random sampling, rather than systematic samplingtechniques, especially in regard to the open oceans.Moreover, means of several measurements from a singlelocale are typically presented, without reference to thepopulation variance. It was impossible, therefore, to testhypotheses about means using t-tests or their non-parametric equivalents. Instead, means of means arereported here, and compared across the two categoriesof coastal surface waters and open ocean surface waters.Such procedures do not enable statistically valid con-clusions, but rather provide a broad qualitativeimpression of the state of the oceans.

Volume 26/Number 3/March1993

It is to be emphasized, however, that the databaseemployed here is identical to that employed byGESAMP. Therefore, although the procedures usedhere cannot satisfy normal standards of scientific con-fidence, they nonetheless permit a critical evaluation ofGESAMP's conclusions. Data are summarized in theform of histograms comparing the mean concentrationof specific contaminants in coastal and open oceansurface waters, in various regions of the world (Figs 1-9).

ResultsPolychlorinated biphenyls (PCBs) are more generally

concentrated in open ocean waters than in coastalwaters (Fig. 1). Average concentrations range from lessthan one-tenth of a nanogram per litre of sea water water(Kanechlor in the Antarctic) to more than 75 nanogramsper litre (Aroclor 1254 in the Atlantic and Pacific). Thehigher concentrations in the open ocean presumablyreflect the atmospheric input of PCBs to ocean water(GESAMP, 1989).

Radioactive wastes may likewise be more concen-trated in the open ocean environment, although theseare not covered in the GESAMP report. This conclusionis based on the historical fact that the deliberatedumping of radioactive wastes at sea (IAEA, 1990),combined with the loss of several naval nuclear reactorsat sea, have deposited far more radioactivity in the openocean than has entered coastal waters (Davis, 1990;Davis and Van Dyke, 1990). Radioactive wastes, how-ever, to the extent that they have escaped containment,have relatively high sediment binding (distribution)coefficients, which restricts their concentration in oceansurface water.

The global distribution of DDT and its metabolitesfollows a slightly different pattern from that of PCBs.Where comparisons are possible between concentra-tions in coastal and open ocean waters, the values areapproximately equal (Fig. 2). In approximately half thecomparisons, concentrations are greater in the openoceans than in coastal waters. By far the greatest levelsare found in the northeast Atlantic, where concentra-tions of about 500 nanograms per litre are substantiallygreater than concentrations in comparable coastal

Phenoclcr D-P510

8

6

4

2

0 O/GMed./France

Fig. 1

zones. As a generality, concentrations are larger in theopen ocean in regions of the world where terrestrialuse of DDT's has been curtailed (e.g. the northernhemisphere), while coastal concentrations are greater inregions where terrestrial use continues (e.g. the southernhemisphere). This distribution pattern presumablyreflects the input of this contaminant to the ocean fromriverine sources. Remote regions, where terrestrial DDTuse has never occurred (e.g. the Antarctic), also showhigher concentration in open oceans, though generallynot as high as in industrialized regions, suggestingincomplete ocean mixing. Assuming a relatively longresidence time for DDT in comparison with mixingrates, existing concentrations in these remote regionsmay be expected to increase over time as mixingbecomes more complete.

A similar analysis for chlorinated hydrocarbons illus-trates that concentrations in sea water are generallyslightly (less than a factor of two; north Pacific, north-west Atlantic) to substantially (approximately an orderof magnitude; Mediterranean Sea) higher in coastalwaters (Fig. 3). Only in remote regions, such as theAntarctic, are concentrations of chlorinated hydro-carbons higher in open ocean waters.

The final class of pollutants examined here, namelythe heavy metals, includes cadmium, lead and mercury.Cadmium is generally moderately to substantially moreconcentrated in coastal zones than the open ocean (Fig.4). Quantitative analysis shows, however, that over theentire globe the concentration is only about 2.3 timeshigher in coastal waters than in the open ocean.Inspection of the sources of cadmium contamination(Fig. 5) suggests that much of it originates from localcoastal "hot spots" such as the New York Bight in theAtlantic, and various North Sea point sources. Whenthese local hot-spots (defined, for this analysis, asregions in which the concentration exceeds twice themean value) are eliminated from the means, the concen-tration ratio (coast/ocean) drops to 1.6. In other words,the concentration of cadmium in coastal waters is lessthan twice as great as in the open ocean waters. Thesedata are consistent with a dynamic source model inwhich cadmium originates largely from land basedsources and is introduced into coastal regions, largely byrun-off but also (potentially) atmospheric input. Con-sidering that the open ocean surface waters represent a

Aroclor 1260 Aroelor 1254 Kanechlor18 0.10

15 75 0.08

12 600.06

9 450.04

6 303 0.02

15

0 0 0.00 O/GO/G O/GA/A.P AP/AP AntlAnt

Mean concentrations of PCBs in ocean (0) and coastal (C) surface waters (nanograms perlitre). Abbreviations: Med., Mediterranean Sea; A, Atlantic; P, Pacific; Ant, Antarctic.Data are means of means presented in tables in Fowler (1990). Numbers above each barrepresent the 'sample' size (where individual samples are, in some cases, mean values ofseveral measurements).

129

vastly greater volume than coastal surface waters, thesedata demonstrate that the bulk of cadmium has found itsway to the open seas.

Lead exhibits a similar pattern to cadmium. Leadconcentrations are generally higher in coastal waters,although occasional comparisons reveal greater concen-tration in the open ocean (Fig. 6). The global ratio ofcoastal to open ocean concentration is approximately

Marine Pollution Bulletin

2.4, but when localized hot-spots are eliminated (Fig. 7),the ratio drops to 1.6.

Finally, mercury concentrations present a similarpattern to cadmium and lead. Concentrations are gener-ally, but not always, higher in coastal waters, but thedifference is generally moderate (Fig. 8). Elimination oflocal hot-spots again yields roughly equivalent concen-trations in coastal and open ocean waters (Fig. 9).

(1~2

(J~ ~ ~ 1A,NE

,...-'''T-- 600 ()500~ I~ ~ :200

a I G

1l0

0:30.061 0.0,./'? 0.04

A Mid 1.Jr;r0.02

::. AS iJ:: lndian WO/G 04 ~

02 0.02 ~O ~:~;I00 0-001 0.02

~ 0010.00 000O/G

5[fP'N P,NE: 20 L2 10

~ 01 G 0

O/G

a

North A,N

o

Fig. 2

a

Mean concentrations of DDT and its metabolites (nanograms per litre) in open ocean (0)and coastal (C) surface waters. Abbreviations: 0, concentration in ocean surface waters;C, concentrations in coastal surface waters. Histograms are positioned in the approximateregions where measurements were made. Further abbreviations: N, north; Chin, ChinaSea; NW, northwest; S. south; otherwise as in Fig. 1. Data are means of means presented intables of Fowler (1990).

.u~ A'

~

O'~: ~ 0 O/G

0.04

0.02

0.00"-a I G '"--"'-

~'~JjP'N0.60.4

0.2

0.0O/G

~'~rIA'S0.60.40.2

0.0

a

Baltic

AntarcticO.8r----

a

(JIG o

Fig. 3 Mean concentrations of chlorinated hydrocarbons (nanograms per litre) in open ocean (0)and coastal (C) surface waters. Otherwise as in Fig. 2.

130

Volume 26/Number 3/March 1993

Discussion state of the oceans suggests that ocean contaminantscan be divided into three broad categories. The firstcategory, including PCBs and arguably radioactivewastes (IAEA, 1990), is more concentrated in the open

Re-analysis here of the admittedly incomplete databased employed by GESAMP in its recent report on the

ole

~[io

ole

Tio 0

a

a

Fig.4 Mean concentration of cadmium (nanograms per litre) in open ocean (0) and coastal (C)surface waters. Otherwise as in Fig. 2.

Open Ocean Surface Waters Coastal Surface Waters

300

250

200 M M M MW W

ec co 1'0 ....0> (]> co '" (]>

'"::. ::. 0> (]> z;150 .2 .2 o u 2 ~

C c :.c ~ :a ·021 21 o 0 ro '"ro '" Q. Q.

100 :;: :;: Q. Q.

;;; ~ ~ ;;;~ '" ro '";t '" ~ ~

50 s: s: s:s: t: t: t:t: 0 0 00 z z zz

0

50

40

30

20

10

aFig. 5

300~

18 2500> E

.2'<Ii s: N"'"eo t: t: CD 200 -~~~ ~ "_0

GO CD ~->s: 0> ~ 6 ::,..g ~

CD ,,~00 ;,; ..• .;, 0> 150 ~~z~ 0> 00

00 ::::. g<z, en ~ c ~~N· ::::. '" ~'i3co co •• 100 --,~en ~ •• 8 uSJ::. (j) (j)

0 s: c "0

~ t: '"0 '6 50al Z .£

0

50

40

30

20

10

0Histograms showing concentrations of cadmium (nanograms per litre) in open ocean (leftpanels) and coastal (right panels) surface water, by region, illustrating locally highconcentrations (point sources or 'hot spots). Lower graphs are of the same data in theupper graphs but on an expanded scale. Data are means of means presented in tables ofFowler (1990).

131

ocean than in coastal regions, reflecting their input fromatmospheric sources (PCBs) or past dumping at sea(radioative wastes). The second category of contami-nants, including DDT and its metabolites, is approxi-

Marine Pollution Bulletin

mately equally concentrated in the open ocean andcoastal waters, suggesting inputs from land sourcesthat have reduced over time. And the third class ofcontaminants, including chlorinated hydrocarbons and

~f1o

O/C

:~~o 01 C

a

Baltic

Fig.6 Mean concentration of lead (nanograms per litre) in open ocean (0) and coastal (C)surface waters. Otherwise as in Fig. 2.

Open Ocean Surface Waters500 e450 co

CD ~ ~ (ij"co co400 '" ~ M..; ~~ co ~,:,

M350 co ~ ~ ..: ~'"

coM ~

coa.M e '"300 i co •• J; ~co ~ ..

o '" c: ..; co250 ~ e co '" '"o '"~ .!l .s ;;: a; z,

c:( :e :e c ;r ii co200 c c '" co •• ~ "•• " a. .. e 8~ Q. Q. '" ::E If)

150 co "in z, .t:

" "in " -c:s: .. " ~ ~100 t: .. s:~ J:;

~ et:0 <50 z If)

0

100,-------------------------------,

80

60

40

20

o

'"~

i-g

Coastal Surface Waters "c'" u:

500 1;'--0:)_'0CD '" ~'" ~450 e ., :::. (!I

M.s .• '""' i c .• ~ •• '" '"•...~

. .. ~400 ~ ~ e '": 0. u "E " ~: .; '" ~ .• ~350 <i .!1 0 ~" . eX c '.ij ell u:- 8- ~"' 0 ~;n- :;; 0 .g ·E ell c ~ '"300 '" "' ~ s: ~ '" ::i ~

'" ~ iD 1! "iii ~ E .• 1; ~M z, ~ ;n- e ~ CD l ~ .,;250 ., ., :E a; a:. i' eX l'" :E '" '" .; t: ~.!l' .!l' s: C "' . ~ e t-o; '" < .• s:

CD '§ u ~~ '" '"200 u ~ ~ ~ . '" •• t: C

C M Iii c .!1 . s ..! .!1 ~ <i .3 .~ ~ ~ ~ '"

.;~!! C I;' .. o 0. ~ M .150 « .'! c 0 s: '""8 « CD ~ . 't; ~ '" 15 s: c~ g- o Z t: .'l!100 z o ,! 0. ::; '" 0 ~z

500

100

'I I•• I. II

80

60

40

20

oFig.7 Histograms showing concentrations of lead (nanograms per litre) in open ocean (left

panels) and coastal (right panels) surface waters, by region, illustrating locally highconcentrations (point sources or 'hot spots'). Otherwise as in Fig. 5.

132

Volume 26/Number 3/March 1993

the heavy metals, is generally somewhat more concen-trated in coastal waters, although elimination ofespecially large point sources of contamination Chotspots') in coastal waters nearly eliminates the difference

with the open oceans. The distribution pattern of thethird class of contaminants presumably originates fromcontinuing coastal inputs.

Although the data reviewed here do not enable statis-

Ber

j~~~r.0.0 ole

North Baltic Arc4 3

0.4

0.2

0.0 ole

lio ole

~r.io ole

a \0.5fjP'SW0.4 I0.3

D 0.2

.0.1

0.0a

Fig.8 Mean concentrations of mercury (nanograms per litre) in open ocean (0) and coastal (C)surface waters. Otherwise as in Fig. 2.

Open Ocean Surface Waters Coastal Surface Waters100~----------------------------.908070605040302010o~ aL~~ ~U.u.~~

toeaOl

;;:- Gl100 eo '"Ol Ol

90 :E .2

80 co ~iii

~;( ~70 r::- >- w enZ ~60 '" JOl

'" ~ .2z; Z t: ~50 Gl '" "@ '"co Ol a. CD;:;- Ol e- Wr-,

40 ec >- e z SOl >- ~ ::>~ '" ;;; .,;30 ~ '" '" <:iii ::>

iii cZ20100

10

8

6

4

2

0

z; ii'>",-a>., '"(/) z;.c: Cd~ <l>

Z (/)

>. .c:

"'" 0to ~ z

18 cc 0en Zen z; 'Ez;

.2ii'> .2.eo u.

~ ene<II to II?a.

ec '" '"Ol ~ >

~

8

10~----------------------------,

6

4

2

oFig.9 Histograms showing concentrations of mercury (nanograms per litre) in open ocean (left

panels) and coastal (right panels) surface waters, by region, illustrating locally highconcentrations (point sources or 'hot spots'). Otherwise as in Fig. 5.

133

ticallyvalid conclusions, the data are identical to thoseon which GESAMP based its summary conclusions. Thedata do not support the unqualified 'dirty coast cleanseas' hypothesis. Instead, they indicate that the openoceans are approximately as contaminated as coastalwaters. Therefore, if coastal waters are considered to beseriously impacted, as stated by GESAMP, then so alsoare open ocean waters. Coastal zones suffer greaterpollution from erosion, turbidity, direct sewage dis-charge and riverine inputs; but by the same token, theopen oceans may suffer more from deliberate dumping,which, with the exception of dredged materials, isgenerally not undertaken in coastal waters; and fromatmospheric inputs, if ony because the receptive surfacearea of the open ocean is far greater.

The levels of contamination in open ocean surfacewaters revealed by the available data merit moredetailed attention for several interacting reasons: 1. 90%of biological productivity occurs there, and this produc-tivity is concentrated in surface waters; 2. The concen-trations and biological impacts of literally hundreds ofocean contaminants are unknown; 3. Concentrationsrecorded in the open ocean (as high as hundreds ofnanograms per litre) appear greater than those known,for some pollutants (e.g., tributyl tin), to produce signifi-cant biological defects (1-2 nanograms per litre; Smith,198]; Goldberg, 1990); 4. The additive, or interactive,effects of ocean stressors, such as the combination ofincreased ultraviolet radiation from ozone depletionwith existing levels of contamination, are largelyunknown; 5. The surface micro layer is lipophilic owingto the concentrations of biological membrane, andhence concentrates organic contaminants such as PCBs;6. Concentration factors in living unicellular organismscan be as high as I00 000 times above the ambient waterconcentration; 7.Developmental events crucial to mostocean life forms occur within the surface micro layer; 8.Approximately one-third of ocean pollution is trans-mitted to ocean waters through this surface microlayerfrom the atmosphere (GESAMP, 1989, 1990); 9.

Marine Pollution Bulletin

Concentrations of contaminants in the surface micro-layer have generally not been measured; and 10.Potential damage from these contaminants may occur ina binary or discontinuous (threshold) mode and beirreversible owing to long time lags.

The GESAMP report on the state of the oceansacknowledges the general inadequacy of the availabledatabase, and especially the paucity of data availablefor the open ocean. The GESAMP report also warnsof the need for concerted international action withinthe next decade if serious and potentially irreversibledegradation of the oceans is to be averted. The presentre-examination of available data supports that recom-mendation, but suggests further that the oceanic eco-system may presently be more vulnerable than suggestedby GESAMP's summary conclusions.

Davis, W. J. (1990). Global aspects of marine pollution policy: The needfor a new international convention. Mar. Po!. 14, 191-197.

Davis, W. J. & Van Dyke, J. M. (1990). Dumping of decommissionednuclear submarines at sea. A technical and legal analysis. Mar. Pol. 14,467-476.

Fowler, S. W. (1990). Concentration of selected contaminants in water,sediments and living organisms. In UNEP: Technical annexes to thereport on the state of the marine environment. UNEP Regional SeaReports and Studies No. 11412. UNEP, 1990, pp. 209-230.

Goldberg, E. D. (1990). Selected contaminants: Tributyltin andchlorinated hydrocarbon biocides.ln UNEP. Technical annexes to thereport on the state of the marine environment. UNEP Regional SeasReports and Studies No. 11412. UNEP, 1990, pp. 209-230.

GESAMP (1989). The Atmospheric Input of Trace Species to the WorldOcean, 1989, Rep. Stud. GESAMP (38).

GESAMP (1990). The state of the marine environment. Rep. Stud.GESAMP (39).

International Atomic Energy Agency. Inventory of Radioactive MaterialEntering the Marine Environment. Part I: Sea Disposal of PackagedLow Radioactive Waste. Issued in draft form in Sept. 1990 asdocument LDC/3/INF.23 by the London Dumping Convention.Vienna: IAEA, and London: International Maritime Organization(1M0).

Smith, B. S. (1981). Male characteristics in female snails caused by anti-fouling bottom paints. 1. Appl. Toxicol. 1,22-25.

UNEP (1990). Technical annexes to the report on the state of the marineenvironment. UNEP Regional Seas Reports and Studies No. 11412.UNEP, 1990 (Volumes 1 and 2).

134