Multidecadal increase in plastic particles in coastal ocean sediments

Microplastics in ocean sediments increased exponentially from 1945 to 2009, serving as a geological proxy for the Anthropocene.


This PDF file includes:
Supplementary Text Fig. S1. Santa Barbara Basin bathymetry and sampling locations.  Table S1. FTIR spectroscopy survey of box core. References (33-40)

Obtaining the sediment core
We assume that this single sediment core is representative of the variability in deposition rate over the basin. The close alignment of stratigraphic events between the Cal-ECHOES sediment cores and SPR0901-06KC, the most recently and accurately dated SBB sediment core, indicates similar conditions among cores (6, 24).

Imaging and dating the sediment core
Hendy et al. (2013)(6) and Schimmelmann et al. (2013)(24) used 14 C dates from planktonic foraminiferal carbonate and terrestrial-derived organic carbon from Kasten core SPR0901-06KC to show that accuracy of the traditional varve counting method decreases prior to approximately 1700 AD due to under-counting of varves. The present box core was sufficiently shallow that it did not necessitate this correction.

Identification of plastics via FTIR spectroscopy
Overall, FTIR permitted definitive identification of 53.0% of the subsampled visually identified microplastics to specific plastic polymer type, and confirmation of an additional 34.5% as most likely plastic (Fig. 3, table S1). Column 5, table S1 equals the 53.0% of definitively identified plastics. The additional 34.5% equals Column 4 minus Column 5 values, since every particle identified as most likely plastic is in Column 4, and a subset of those values that were definitively identified is in Column 5.
Subsampling was done by the senior author at the time of FTIR analysis. This was done visually, by removing identified particles by forceps from their sample slides, and placing them on the FTIR. At least 10% of each transect layer was analyzed this way, with every effort to mimic the proportions of plastics collected (i.e., sampling more fibers than fragments, film, or spherical particles due to their higher presence in core, fig. S4). There is most likely some selection bias in the subsampling, where particles were so small that the senior author could not see them or properly transfer them for FTIR analysis. There is also slight overrepresentation of fragments and film particles (table S1, Column 3), because they are easier to analyze via FTIR than fibers.
Some of the particles that could not be definitively identified to plastic type had spectra that looked similar to two plastic types that could not be differentiated. This problem was common with particles that showed the triplicate diagnostic peak of polyvinyl chloride (PVC) and polyethylene terephthalate (PET) and particles that showed the doublet peak of polyethylene (PE) and polystyrene (PS), but could not be differentiated further, either because of the quality of the spectra, particle weathering that occluded differentiating regions of the spectra (33), or because the real-world plastic particles contained additives, colorants, etc. that are known to change the shape of FTIR spectra from pure standards (34). LDPE was differentiated from HDPE by the presence of a small peak at 1377 cm -1 , and if its presence/absence was not clear from the spectra, the piece was recorded as PE (33).
The particles that could not be identified (table S1, Column 6) were often too small or thin to generate good FTIR spectra or they had readings that were not conclusive. Some of those particles were visually recorded at the time of FTIR reading as "thin white fiber" or "reddish black fragment," so they are likely plastic, but this cannot be confirmed. A few particles had clean spectra similar to the spectra of calcium carbonate, aragonite, or clay (table S1, Column 7)(29). These particles were visually recorded at the time of FTIR reading as "filmy fragment," "tiny off-white fragment," "fragment/sediment, looks biological," or "really thin fiber with sediment still stuck to it." It is likely that these particles were organic or sedimentary, or in the last case, had sediment still attached to them. Particles that appeared organic or sedimentary under microscopy were already removed from the plastic enumerations during the secondary proofing of the images by the senior author. However, these particles were removed from the plastic abundance calculations but were not physically removed from the paleontological sorting tray. Thus, these particles that were subsampled and analyzed by FTIR (table S1), and had organic or sedimentary signatures, had already been excluded from the plastic abundance enumerations.

Plastics as contamination
Any plastic before 1945 was treated as contamination due to the low amounts of plastic in production at that time (7). However, this procedure could overestimate contamination, as synthetic plastic was invented in 1907 (7) and did exist in small amounts before World War II (7). But we assume nearly all of these pieces were added during core processing.
The trilaminate bag in which the core was stored was PET, as identified by FTIR, and any contamination from the bag could not be differentiated from other PET unless the fragments were large enough to determine whether they were metallic silver (like the trilaminate bag) or another color (Fig. 3B). However the bag was removed before the core was cut, and so any trilaminate bag contamination would be incidental. Modest contamination was found from the plastic core liner box that was cut with a saw to make the core chronology; small numbers of un-aged, identical fragments of that unique spectrum were found intermittently throughout the core (Fig. 3E; fig. S4B). Those fragments were included in Column 4 and 5 of table S1, because these fragments were known to be plastic and identified to type (type = core liner, because its unique FTIR spectrum is not identical to any published standard). These fragments were obviously contamination, and were thus accounted for in the baseline contamination subtraction.
We attempted to use the spectral signatures of plastic degradation (33) to assess which pieces were aged and thus more likely to be contamination. However, a marked temporal trend in aging was not seen between layers of the core denoting contamination and noncontamination plastics.

Fiber contamination and deposition
The majority of plastics found in the core overall were fibers, including in the  (2011), we hypothesized that most of the contamination would be airborne fibers, as was found. But the fact that the core spans years before and after the prevalence of plastic allows us to subtract the contamination values and determine that there are still numerous fibers buried in the sediment itself, which agrees with earlier literature (2, 35).

Plastic fragments and spheres
Fourteen percent of the particles in the core were plastic fragments, although many more in the post-1945 samples than pre-1945 samples ( fig. S3). Almost no spherical plastics were found in the core, despite the fact that microbeads in consumer products have a diameter of ~400-500 μm and would have been retained in the 104 μm sieve (14, 38).
One of the only perfectly spherical microbeads we found was an un-aged polystyrene microbead in an early sample (circa 1893, before the advent of plastic) and was thus contamination.

Buoyancy of plastics
PE is less dense than seawater and would not logically be found in the sediment; however we found multiple pieces of it. One piece of plastic was also identified as most likely polypropylene (PP), which is also less dense than seawater. However, Browne et al.
(2011)(2) also found polypropylene and polyethylene in sediment samples near areas of sewage and wastewater effluent, and it is likely that our samples contain particles from the effluent and riverine flow off the Santa Barbara coastline. These buoyant plastics could have also sunk to the bottom through fecal or marine snow transport (22, 39), or via biofouling (40). It is also possible that these buoyant pieces were airborne or processing contamination.