Lost at Sea: Where Is Allthe Plastic?

Richard C. Thompson,1* Ylva Olsen,1 Richard P. Mitchell,1

Anthony Davis,1 Steven J. Rowland,1 Anthony W. G. John,2

Daniel McGonigle,3 Andrea E. Russell3

Millions of metric tons of plastic are producedannually. Countless large items of plastic debrisare accumulating in marine habitats worldwideand may persist for centuries (1–4). Here weshow that microscopic plastic fragments and fi-bers (Fig. 1A) are also widespread in the oceansand have accumulated in the pelagic zone andsedimentary habitats. The fragments appear tohave resulted from degradation of larger items.Plastics of this size are ingested by marine organ-isms, but the environmental consequencesof this contamination are still unknown.

Over the past 40 years, large items ofplastic debris have frequently been re-corded in habitats from the poles to theequator (1–4). Smaller fragments, proba-bly also plastic, have been reported (5) buthave received far less attention. Mostplastics are resistant to biodegradation, butwill break down gradually through me-chanical action (6). Many “biodegrad-able” plastics are composites with materi-als such as starch that biodegrade, leavingbehind numerous, nondegradable, plasticfragments (6). Some cleaning agents alsocontain abrasive plastic fragments (2).Hence, there is considerable potential forlarge-scale accumulation of microscopicplastic debris.

To quantify the abundance of micro-plastics, we collected sediment frombeaches and from estuarine and subtidalsediments around Plymouth, UK (Fig.1B). Less dense particles were separatedby flotation. Those that differed in appear-ance to natural particulate material (Fig.1A) were removed and identified withFourier Transform infrared (FT-IR) spec-troscopy (7). Some were of natural originand others could not be identified, butabout one third were synthetic polymers(Fig. 1C). These polymers were present inmost samples (23 out of 30), but weresignificantly more abundant in subtidalsediment (Fig. 1D). Nine polymers wereconclusively identified: acrylic, alkyd,poly (ethylene:propylene), polyamide(nylon), polyester, polyethylene, poly-methylacrylate, polypropylene, andpolyvinyl-alcohol. These have a widerange of uses, including clothing, packag-

ing, and rope, suggesting that the fragments result-ed from the breakdown of larger items.

To assess the extent of contamination, a fur-ther 17 beaches were examined (Fig. 1B). Similarfibers were found, demonstrating that microscopicplastics are common in sedimentary habitats. Toassess long-term trends in abundance, we exam-ined plankton samples collected regularly sincethe 1960s along routes between Aberdeen and theShetlands (315 km) and from Sule Skerry to Ice-

land (850 km) (7) (Fig. 1B). We found plasticarchived among the plankton in samples back tothe 1960s, but with a significant increase in abun-dance over time (Fig. 1E). We found similar typesof polymer in the water column as in sediments,suggesting that polymer density was not a majorfactor influencing distribution.

It was only possible to quantify fragments thatdiffered in appearance from sediment grains orplankton. Some fragments were granular, butmost were fibrous, �20 �m in diameter, andbrightly colored. We believe that these probablyrepresent only a small proportion of the micro-scopic plastic in the environment, and methods arenow needed to quantify the full spectrum of ma-terial present. The consequences of this contami-nation are yet to be established. Large plasticitems can cause suffocation and entanglement anddisrupt digestion in birds, fish, and mammals (3).To determine the potential for microscopic plas-tics to be ingested, we kept amphipods (detriti-vores), lugworms (deposit feeders), and barnacles(filter feeders) in aquaria with small quantities ofmicroscopic plastics. All three species ingestedplastics within a few days (7) (fig. S1).

Our findings demonstrate the broad spatialextent and accumulation of this type of contam-ination. Given the rapid increase in plastic pro-duction (Fig. 1E), the longevity of plastic, andthe disposable nature of plastic items (2, 3), thiscontamination is likely to increase. There is thepotential for plastics to adsorb, release, andtransport chemicals (3, 4). However, it remainsto be shown whether toxic substances can passfrom plastics to the food chain. More work isneeded to establish whether there are any envi-ronmental consequences of this debris.

References and Notes1. P. G. Ryan, C. L. Moloney, Nature 361, 23 (1993).2. M. R. Gregory, P. G. Ryan, in Marine Debris, J. M. Coe,

D. B. Rogers, Eds. (Springer, Berlin, 1996), pp. 48–70.3. J. G. B. Derraik, Mar. Pollut. Bull. 44, 842 (2002).4. E. J. Carpenter, S. J. Anderson, G. R. Harvey, H. P.

Miklas, B. P. Bradford, Science 178, 749 (1972).5. J. B. Colton, F. D. Knapp, B. R. Burns, Science 185, 491

(1974).6. P. P. Klemchuck, Polym. Degrad. Stab. 27, 183 (1990).7. Materials and methods are available as supporting

material online on Science Online.8. We thank C. Hoare, R. Ticehurst, G. Mandair, and F.

Birembaut for help with sample collection and anal-ysis. Supported by the Leverhulme Trust, UK.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/304/5672/838/DC1Materials and MethodsFig. S1References and Notes

10 December 2003; accepted 10 February 2004

1University of Plymouth, PL4 8AA, UK. 2Sir AlisterHardy Foundation for Ocean Science, Plymouth, PL12PB, UK. 3University of Southampton, SO17 1BJ, UK.

*To whom correspondence should be addressed. E-mail: rcthompson@plymouth.ac.uk

Fig. 1. (A) One of numerous fragments found amongmarinesediments and identified as plastic by FT-IR spectroscopy. (B)Sampling locations in the northeast Atlantic. Six sites nearPlymouth (▫) were used to compare the abundance of mi-croplastic among habitats. Similar fragments (●) were foundon other shores. Routes sampled by Continuous PlanktonRecorder (CPR 1 and 2) were used to assess changes inmicroplastic abundance since 1960. (C) FT-IR spectra of amicroscopic fragment matched that of nylon. (D) Microplas-tics were more abundant in subtidal habitats than on sandybeaches (*, F2,3 � 13.26, P � 0.05), but abundance wasconsistent among sites within habitat types. (E) Microscopicplastic in CPR samples revealed a significant increase inabundance when samples from the 1960s and 1970s werecompared to those from the 1980s and 1990s (*, F3,3 �14.42, P� 0.05). Approximate global production of syntheticfibers is overlain for comparison. Microplastics were also lessabundant along oceanic route CPR 1 than along CPR 2(F1,24 � 5.18, P � 0.05).

BREVIA

7 MAY 2004 VOL 304 SCIENCE www.sciencemag.org838