Marine Pollution Bulletin 88 (2014) 249–254
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
The effect of fine-scale sampling frequency on estimates of beach litter accumulation Peter G. Ryan a,⇑, Annerie Lamprecht b, Debbie Swanepoel a, Coleen L. Moloney b a b
Percy FitzPatrick Institute of African Ornithology, DST/NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa Marine Research Institute and Department of Biological Sciences, University of Cape Town, Rondebosch 7701, South Africa
a r t i c l e
i n f o
Article history: Available online 16 September 2014 Keywords: Stranded debris Plastics Sample rate Accumulation Meta-analyses South Africa
a b s t r a c t The effect of sampling frequency on estimates of the rate of litter accumulation was determined for two South African sandy beaches. After initial cleaning, all manufactured items >10 mm diameter were collected in alternating bouts of daily or weekly cleanups. Daily sampling collected 2.5 (range 2.1–3.4) times more litter items than weekly samples and 1.7 (1.3–2.3) times more litter by mass. Low density items such as foamed polystyrene showed a greater differential (4–5 times more items from daily sampling), presumably due to faster turnover of lightweight litter items. Variation in weekly samples was not consistently less than daily estimates, suggesting that less frequent samples only partly integrate short-term fluctuations in litter dynamics. Researchers using beach accumulation data to infer trends in nearshore marine litter, or to assess the efficacy of litter mitigation measures, need to ensure consistency in sampling frequency. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Plastic and other drifting artefacts are major marine pollutants, and considerable efforts have been made to prevent the release of plastic articles into the environment (e.g. Coe and Rogers, 1997). Assessing the efficacy of measures to reduce marine debris at sea is complicated by its great spatial heterogeneity and the high cost of at-sea surveys (Ryan et al., 2009). As a result, much of what we know about marine litter is inferred from surveys of litter stranding on beaches (e.g. Derraik, 2002; Barnes et al., 2009), and major regional initiatives to assess trends in marine debris rely heavily on beach monitoring (Sheavly, 2007; OSPAR Commission, 2007). Initial studies of beach litter reported standing stocks of accumulated litter, which can show gross changes in the abundance and distribution of litter (e.g. Ryan and Moloney, 1990), but are not necessarily a good indicator of the abundance of litter in adjacent coastal waters (Escardó-Boomsma et al., 1995; Ryan et al., 2009). Other factors that affect the amount of litter on beaches include the physical environment (beach structure, wave action and local currents, which influence both stranding and removal rates); beach dynamics (burial or exposure of litter); proximity to urban centres (which influences exposure to land-based sources
⇑ Corresponding author. Tel.: +27 21 6502966; fax: +27 21 6503295. E-mail address: [email protected] (P.G. Ryan). http://dx.doi.org/10.1016/j.marpolbul.2014.08.036 0025-326X/Ó 2014 Elsevier Ltd. All rights reserved.
such as the number of beach visitors and urban run-off); and formal or informal cleanup efforts (Merrell, 1980; Bowman et al., 1998; Williams and Tudor, 2001). Also, increases in the standing stock of beach litter may reflect long-term accumulation rather than a change in the amount of debris at sea (Ryan et al., 2009). As a result, estimates of litter accumulation rates made by repeatedly cleaning the same stretch of beach currently are best practice for using beach surveys to assess marine litter trends (e.g. Sheavly, 2007; OSPAR Commission, 2007; Ryan et al., 2009; Ribic et al., 2010, 2012). One problem with litter accumulation estimates is that they might be sensitive to the frequency with which litter is sampled because of differences in litter turn-over rates (Ryan et al., 2009). Dixon and Cooke (1977) showed that the weekly retention rate of marked bottles on beaches in Kent, UK, varied among beaches and was influenced by the type of bottle, because plastic bottles remained ashore longer than glass bottles. Subsequent studies have confirmed different retention rates depending on beach structure and the type of litter (Merrell, 1980; Garrity and Levings, 1993; Bowman et al., 1998; Kataoka et al., 2013). EscardóBoomsma et al. (1995) differentiated the amount of litter arriving on a beach (the loading rate) from the amount of litter that accumulates per unit of time (net accumulation rate). However, this distinction has been largely ignored by subsequent studies, which have reported accumulation rates at sampling frequencies ranging from daily (Eriksson et al., 2013), every two weeks (Williams and
250
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Tudor, 2001), monthly (Corbin and Singh, 1993; Thornton and Jackson, 1998; Sheavly, 2007), quarterly (OSPAR Commission, 2007), every six months (Slip and Burton, 1991) to annually (Merrell, 1984; Edyvane et al., 2004). Some studies have sampled at different frequencies in different seasons (e.g. Walker et al., 1997; Ivar do Sul et al., 2011). Sheavly (2007) reported no difference in accumulation rate for samples collected at different intervals, but Ribic et al. (2010, 2012) found that frequent sampling (614 days) and very infrequent sampling (>180 days) affected debris counts relative to their 28-day standard sampling interval. Ribic et al. (2010, 2012) failed to report how these changes in sampling frequency affected litter counts, but the inference was that too frequent sampling reduced litter counts, and very infrequent samples increased litter counts. However, Ryan et al. (2009) included results from an unpublished MSc dissertation (Swanepoel, 1995) to suggest that summing daily accumulation samples to obtain weekly data yielded up to sixtimes more litter items than weekly sampling at the same beaches. More recently, Eriksson et al. (2013) showed that daily sampling at sub-Antarctic Macquarie Island increased litter accumulation rates almost ten-fold compared to monthly sampling, although surveys took place in different years. These findings suggest that sampling frequency, at least at fine temporal scales, has a major impact on estimates of litter accumulation rates. In this paper we report the findings from Swanepoel’s (1995) work in more detail, and supplement them with additional sampling. We compare the amount (number and mass) of litter collected daily with weekly accumulations collected at the same sites to evaluate the extent to which more frequent sampling increases estimates of the amount of litter arriving on beaches. We also assess whether less frequent sampling reduces the variability in beach litter accumulation data by integrating fine-scale variations (given that processes such as weather, wave action and tidal patterns might be expected to affect litter arrival and exhumation rates at fine temporal scales). Our findings have relevance for the design of beach litter monitoring programmes, and are especially important for meta-analyses of beach litter accumulation data collected at different sampling frequencies.
2. Materials and methods The study took place at two beaches in Table Bay, South Africa (Fig. 1), in the Cape Town metropolitan area. The beaches were chosen to have similar physical structures but to be different distances from the urban centre and to have contrasting ease of access and user profiles. Milnerton Beach is a popular recreational beach 10 km from the city centre, used for walking, surfing, swimming and other recreational activities, and attracts large numbers of visitors daily, especially in summer (November–March). Koeberg Nature Reserve is 25 km farther north (35 km from the city centre) and the study beach is closed to visitors. Both are exposed, west-facing, fine-grained sandy beaches with gentle to moderate beach profiles (10–15° slopes), and both are backed by steep, vegetated dunes which limit shoreward migration of litter. Neither study site was cleaned by local authorities during the surveys, but adjacent beaches at both sites were cleaned: Milnerton was cleaned by municipal workers weekly (1994 and 1995) or daily (2012), whereas Koeberg was cleaned by volunteers every four months. We could not control informal cleaning efforts by members of the public. At each beach, a study area was demarcated, divided into ten equal subsections along the beach (Fig. 1). In 1994 and 1995, 500 m of beach was sampled (ten 50-m subsections) at each beach, and the same site was sampled at Koeberg in 2012. However, at Milnerton only the southern 250 m surveyed in the 1990s was sampled in 2012 (ten 25-m subsections) because the marked
increase in the density of litter prevented both beaches being sampled in full on the same day. Prior to each experiment, a team of volunteers removed all accumulated macro litter (articles >10 mm diameter) between the water line and the dune vegetation (including any litter visible in the vegetation) from each study area, and from adjacent buffer zones approximately 25 m wide at each end of the study sections. The following day (24 h, or two tidal cycles later), sampling of ‘new’ litter commenced at both sites. Teams of three to six observers worked systematically along each beach, collecting all litter items in each of the ten subsections. Each subsection was searched until no items had been found for several minutes, with searches of each section taking roughly 5–15 min, depending on the amount of litter present. In 1994, daily sampling continued for 15 days (18 October–1 November), and was followed by one sample after 4 days, one after 6 days and four weekly samples, and then another 14 daily samples (9–23 December). In 1995, the initial cleaning took place on 26 June, with a weekly sample on 3 July, 14 daily samples to 17 July, and then a final weekly sample on 24 July. In 2012, daily sampling continued for 10 days (2–11 October), followed by three weekly samples (18 and 25 October and 1 November), 10-days of daily sampling (2–11 November), four weekly samples (18 and 25 November and 2 and 9 December), and then a final 10 daily samples (10–19 December). There was no consistent difference in tidal phase (neap or spring) between daily and weekly sample periods. All litter samples were returned to the University of Cape Town to be sorted and counted. Each debris item was identified as far as possible and categorised by type of material (plastic, metal, glass, cigarette butts, etc.). Wood items were included if they were ‘worked’ rather than natural wood. Plastic items were further subdivided by function: packaging and other single-use items, user items including fishing gear, and plastic fragments of uncertain provenance. Foamed polystyrene items (including cups and fast-food trays as well as packing chips and moulded packaging) were placed in a separate category because of their lower density than most other litter items. Articles were dried and cleaned of sand before weighing. In 1994, the entire sample was weighed to the nearest 25 g with a Pesola spring balance, but large pieces of timber and other very large litter items were not weighed. No mass data were recorded in 1995, but in 2012 each litter category was weighed separately to the nearest 0.1 g on an electronic top-pan balance. Litter samples were kept separate for each 50- or 25-m subsection of beach and data compared among subsections to check for consistent differences in fine-scale accumulation of litter. Such differences were detected, presumably linked to local currents, wave action and resultant beach structure (Bowman et al., 1998), but there was no tendency for the end subsections at each beach to have the highest accumulation rates, suggesting that lateral drift from adjacent areas outside the sampling area was minimal (data not shown). Thereafter, samples were pooled for each beach and accumulation data were summarised as the number or mass of litter items recorded per metre of beach per day, calculated as the litter yield divided by the length of beach sampled (m) and by the number of days since the last collection (days). For comparative purposes, the rate estimated from the 6-day accumulation sample in 1994 was combined with the weekly samples, but the 4-day accumulation sample was discarded. The effects of daily versus weekly sampling on the accumulation of litter on the beaches was tested with a general linear model (GLM), with beach and year as covariates: 1 Accumulated litter ðm1 d Þ ¼ eb0 þ bweekly=daily þ bbeach þ byear þ e
where b0 is a constant; bweekly/daily, bbeach and byear are parameter sets describing the effects of the predictor variables (sampling
251
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
10 km Koeberg
Blouberg
Robben Island
Table View
Table Bay Rietvlei
Milnerton
Black River Cape Town
Fig. 1. Table Bay, South Africa, showing the location of the two study sites at Milnerton and Koeberg, each comprising 10 subsections.
frequency, beach and year) on litter accumulation, and e refers to the residuals (there were no significant interaction effects). The model was applied to all litter items and also to each major category of litter. Where there were zero values in the data set (for nonplastic items), a small constant (equal to 10% of the mean abundance) was added to each data point before a logarithmic transformation was applied to reduce heteroscedasticity and normalise the data (both checked by residual analysis). Coefficients of variation corrected for sample size [=SD/mean ⁄ (1 + (1/4n))] were used to compare relative variability in the number and mass of litter items recorded in daily and weekly samples from the same beach/ year. However, there were too few weekly samples in July 1995 to compare their variation with daily samples.
3. Results In October–December 1994, the 34 accumulation samples at both beaches comprised 38,155 litter items weighing at least 151.1 kg. In July the following year 39,963 items were sampled in only 16 collections, and 37 collections in 2012 yielded 199,303 items weighing 257.8 kg. Plastic articles dominated the samples
numerically from both beaches in all three years, comprising 86.5% by number in 1994, 89.3% in 1995 and 93.5% in 2012. The contribution of plastics to the mass of litter was only recorded in 2012, when they comprised 58.9% overall, with wood (21.9% of total mass) and glass (12.8%) dominating the mass of non-plastic items. Plastics were more dominant at Koeberg (95.3% by number in all years and 62.0% by mass in 2012), which is farther from the city centre and has many fewer visitors than Milnerton (91.3% by number and 57.7% by mass). After plastics, wood contributed most to the mass of litter at both Koeberg (30.4%) and Milnerton (18.4%), but glass items were more important at Milnerton (16.9% of total litter mass) than at Koeberg (2.7%). The GLM indicated that estimates of daily accumulation rates of all litter items varied with sample site (F = 335, df = 1, 169, p < 0.00001), year (F = 81, df = 2, 169, p < 0.00001) and sampling frequency, with daily sampling providing consistently greater estimates than weekly sampling (F = 32, df = 1, 169, p < 0.00001; Table 1). On average, after accounting for the effects of beach and year, daily samples captured 2.5 times more litter overall than weekly samples (Fig. 2). This factor was consistent (1.8–2.5) for all broad litter categories apart from foamed polystyrene (4.2), which had the largest factor of all litter types in five of six beach-year
252
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Table 1 The effect of sampling frequency (daily versus weekly) on estimates of litter accumulation rates (number of items per metre of beach per day) at two South African beaches in three years. ‘Factor’ is the number of times the rate estimated from daily samples exceeds that from weekly samples (=daily rate/weekly rate). Site and litter type
1994
1995
2012
Day n = 29
Week n=5
Factor
Day n = 14
Week n=2
Factor
Day n = 30
Week n=7
Factor
0.40 0.71 0.15 0.05 1.31
0.22 0.11 0.04 0.01 0.39
1.8 6.7 3.3 4.2 3.4
1.35 0.37 0.47 0.27 2.46
0.60 0.04 0.13 0.10 0.86
2.2 10.5 3.7 2.7 2.8
4.32 1.38 1.54 6.26 13.50
1.63 0.32 0.89 2.27 5.11
2.6 4.3 1.7 2.8 2.6
Cigarettes Wood Other non-plastics All non-plastics
0.16 0.04 0.04 0.24
0.05 0.01 0.01 0.07
3.4 4.0 3.1 3.5
0.18 0.06 0.17 0.41
0.02 0.01 0.03 0.06
7.6 8.3 5.6 6.7
0.57 0.13 0.38 1.08
0.21 0.05 0.09 0.36
2.7 2.5 4.1 3.0
All litter
1.55
0.46
3.4
2.87
0.93
3.1
14.58
5.47
2.7
0.11 0.16 0.04 0.02 0.33
0.05 0.07 0.02 0.01 0.15
2.1 2.1 2.3 2.2 2.1
0.38 0.50 0.21 0.12 1.21
0.26 0.10 0.15 0.06 0.56
1.4 5.1 1.5 2.1 2.1
0.75 0.50 0.27 0.48 1.99
0.32 0.09 0.10 0.10 0.61
2.3 5.7 2.7 4.6 3.3
Cigarettes Wood Other non-plastics All non-plastics
0.01 0.01 0.01 0.03
0.00 0.01 0.00 0.01
3.7 2.0 4.9 2.8
0.01 0.04 0.04 0.09
0.00 0.02 0.03 0.06
2.9 1.9 1.2 1.5
0.01 0.02 0.02 0.05
0.00 0.01 0.01 0.02
3.0 2.5 3.3 2.8
All litter
0.35
0.16
2.2
1.30
0.62
2.1
2.04
0.63
3.2
Milnerton Packaging Polystyrene User items Fragments All plastics
Koeberg Packaging Polystyrene User items Fragments All plastics
comparisons (Table 1). The consistent pattern of crude factors when comparing daily to weekly estimates for each site and year (Table 1) provides further evidence for the generality of this effect. Sampling frequency appears to have slightly less of an impact on estimates of the mass of litter accumulating on beaches. The crude mass estimates in 1994 indicated that daily masses at both beaches were 1.5 times greater than estimates based on weekly samples. These values increased to 1.8 (Koeberg) and 2.3 (Milnerton) in 2012, when the average mass of each litter item (1.3 g) was considerably less than in 1994 (4.0 g), despite the fact that large pieces of wood and other very large litter items were not weighed in 1994. Data on how the masses of different categories of litter differed in their response to sampling frequency were restricted to 2012. Plastic litter items tended to have a slightly greater day:week factor (Milnerton 2.5, Koeberg 1.9) than non-plastic items (Milnerton 2.1, Koeberg 1.7). Foamed polystyrene again had the greatest factors (Milnerton 3.2, Koeberg 3.3), although these were lower than the factors for numbers of polystyrene items in 2012 (Table 1). The numbers of litter items accumulating daily in different time periods at each beach varied considerably (Table 2), with an average coefficient of variation of 59%. This was slightly greater than the average weekly coefficients of variation in 1994 and 2012 (49%), but there was no consistent pattern comparing daily and weekly CVs. Daily variation in the mass of litter (64%) also was slightly greater than weekly samples (49%), but again there was no consistent pattern. For example, in 1994, the daily CV in mass was greater than the weekly CV at Koeberg, but weekly variability was greater at Milnerton (Table 2). Coefficients of variation tended to be greater at Koeberg than at Milnerton (Table 2), possibly linked to the smaller volumes of litter at Koeberg. In 1994/95, daily CVs were greatest in winter (July) and decreased through spring (October) to summer (December). However, in 2012 CVs of daily litter counts increased from October to December (but there was no corresponding pattern in CVs of litter masses; Table 2).
4. Discussion Beaches are dynamic systems where litter can be stranded by the sea; dumped by beach visitors; buried and exhumed; exported inland, offshore or along-shore by wind, wave or tidal action; or removed by formal or informal beach cleaning efforts (Ryan et al., 2009). The best way to monitor the persistence of litter items (and, by inference, estimate the scale of these movements and the resultant litter turnover rates) is by mark-recapture studies (e.g. Merrell, 1980; Garrity and Levings, 1993; Bowman et al., 1998; Williams and Tudor, 2001; Kataoka et al., 2013). However, markrecapture studies are logistically challenging on a large scale at urban beaches characterised by large litter loads; most studies to date have only sampled specific litter types (Dixon and Cooke, 1977; Merrell, 1980; Williams and Tudor, 2001). Our study sampled over 280,000 litter items; attempting to mark and follow the fate of even a small subsample of this number would be almost impossible. By alternating daily and weekly beach litter collections we were able to show that more frequent sampling increased the estimate of the amount of litter found on a beach. Our results were consistent across beaches in three different years, sampling in both summer and winter, with little variation in the magnitude of the effect of more frequent sampling. For most litter items, daily sampling resulted in 2–3 times more litter being collected than weekly sampling. This presumably results from rapid turnover of many litter items which are missed in weekly samples because they are either buried or transported off the beach (blown inshore, blown or washed offshore, carried alongshore, or removed by informal beach cleaning efforts by wellmeaning members of the public). Small, lightweight items are more likely to be blown inland or buried, resulting in rapid turnover ( Merrell, 1984; Williams and Tudor, 2001), and hence should have a greater differential between daily and weekly estimates of accumulation rates. This was shown by foamed polystyrene, which is particularly prone to wind dispersal and had more than a fourfold increase in litter accumulation rates based on daily samples.
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Plastics
In addition to differences among litter types, the effect of sampling frequency on estimates of litter arrival rates varies with the magnitude of the difference in sampling frequency. Thus Eriksson et al. (2013) found almost an order of magnitude more litter by sampling daily rather than monthly at beaches on two sub-Antarctic islands. The size of the effect probably also varies between beaches, as a result of differences in turnover rates linked to beach structure (e.g. slope, orientation, exposure) and local conditions (e.g. wind strength and direction). In general, the faster the turnover rate, the greater the disparity between arrival rates estimated at different sampling frequencies (cf. Kataoka et al., 2013). The effect of daily sampling was greater in 2012 than in the 1990s, presumably because the mean mass of litter items in 2012 was lower, increasing turnover rates. Our findings are particularly important for meta-analyses comparing beach litter accumulation data collected at different frequencies. For example, Barnes and Milner (2005) presented broad latitudinal patterns in beach litter density, ostensibly representing annual accumulation rates, but scrutiny of their sources indicate a diversity of sampling approaches from standing stock surveys to seasonal and monthly sampling. Even if analyses control for sampling frequencies, the intrinsic differences in litter turnover rates among beaches (see above) are likely to confound comparisons unless samples are collected at a high frequency (e.g. daily). Unfortunately, daily sampling is costly to implement because it is time and labour intensive. Less frequent, regular sampling (e.g. monthly) of the same site or set of sites is adequate to track long-term changes in litter loads. Such programmes would benefit from short bouts of daily sampling to assess the extent to which infrequent sampling underestimates the loading rates of specific litter items (i.e. to evaluate the proportions of various litter items arriving on a beach that actually accumulate there). This information could be used to develop correction factors to improve the accuracy of monitoring programmes, although turnover rates probably depend on the size as well as type of litter item. Ideally, litter comparisons among beaches should be based on estimates of loading rate rather than accumulation rate.
Non-plastics 0.10
0.8 0.08 0.6 0.06 0.4
Litter accumulation rate (items per metre per day)
0.04 0.2
0.02
0.0
0.00
Packaging
Polystyrene 0.25
0.3 0.20 0.15
0.2
0.10 0.1 0.05 0.00
0.0
Fragments
User items
0.05 0.10 0.04 0.08 0.03 0.06 0.02
0.04
0.01
0.02
Acknowledgements
0.00
0.00
Day
Week
Day
Week
Fig. 2. The effect of sampling frequency (daily versus weekly) on estimates of litter accumulation rates for different categories of beach litter determined by a GLM including data from two Cape Town beaches over three years. Error bars show 95% confidence intervals.
Table 2 Coefficients of variation (CV) in daily and weekly litter collections at two South African beaches in three years. Too few weekly samples were collected in July 1995 to estimate the weekly CV, and there were no mass data for 1995. Year and month
Milnerton
253
Koeberg
Number
Mass
Number
1994 October daily (n = 15) December daily (n = 14) Weekly (n = 5)
48% 24% 27%
59% 44% 71%
1995 July daily (n = 14)
89%
–
130%
–
2012 October daily (n = 10) November daily (n = 10) December daily (n = 10) Weekly (n = 7)
38% 49% 61% 44%
67% 27% 23% 23%
36% 40% 102% 59%
94% 77% 88% 54%
47% 42% 65%
Mass 87% 71% 57%
We thank the numerous volunteers who assisted with litter collections, especially P. Hardcastle, M. Lamprecht, A. Legodi, T. Oosthuizen and J. Oosthuizen. G. Greef, J. Le Roux and H. Westman kindly gave permission to work in the Koeberg Private Nature Reserve. G. Frost (Milnerton Municipality) and J. Kieser (Plastics SA) supplied rubbish bags and assisted with initial cleanups. The National Research Foundation – South Africa and the University of Cape Town – South Africa provided funds for field expenses and bursary support for A.L. and D.S.
References Barnes, D.K.A., Milner, P., 2005. Drifting plastic and its consequences for sessile organism dispersal in the Atlantic Ocean. Mar. Biol. 146, 815–825. Barnes, D.K.A., Galgani, F., Thompson, R.C., Barlaz, M., 2009. Accumulation and fragmentation of plastic debris in global environments. Phil. Trans. R. Soc. B 364, 1985–1998. Bowman, D., Manor-Samsonov, N., Golit, A., 1998. Dynamics of litter pollution on Israeli Mediterranean beaches: a budgetary, litter flux approach. J. Coastal Res. 14, 2588–2597. Coe, J.M., Rogers, D.B., 1997. Marine Debris: Sources, Impacts and Solutions. Springer-Verlag, New York. Corbin, C.J., Singh, J.G., 1993. Marine debris contamination of beaches in St. Lucia and Dominica. Mar. Pollut. Bull. 26, 325–328. Derraik, J.G.B., 2002. The pollution of the marine environment by plastic debris: a review. Mar. Pollut. Bull. 44, 842–852. Dixon, T.R., Cooke, A.J., 1977. Discarded containers on a Kent Beach. Mar. Pollut. Bull. 8, 105–109.
254
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Edyvane, K.S., Dalgetty, A., Hone, P.W., Higham, J.S., Wace, N.M., 2004. Long-term marine litter monitoring in the remote Great Australian Bight, South Australia. Mar. Pollut. Bull. 48, 1060–1075. Eriksson, C., Burton, H., Fitch, S., Schulz, M., 2013. Daily accumulation rates of marine debris on sub-Antarctic island beaches. Mar. Pollut. Bull. 66, 199–208. Escardó-Boomsma, J., O’Hara, K., Ribic, C.A., 1995. National Marine Debris Monitoring Program, vols. 1–2. US EPA Office of Water, Washington, DC. Garrity, S.D., Levings, S.C., 1993. Marine debris along the Caribbean coast of Panama. Mar. Pollut. Bull. 26, 317–324. Ivar do Sul, J.A., Barnes, D.K.A., Costa, M.F., Convey, P., Costa, E.S., Campos, L., 2011. Plastics in the Antarctic environment: are we looking only at the tip of the iceberg? Oecol. Aust. 15, 150–170. Kataoka, T., Hinata, H., Kato, S., 2013. Analysis of a beach as a time-invariant linear input/output system of marine litter. Mar. Pollut. Bull. 77, 266–273. Merrell Jr., T.R., 1980. Accumulation of plastic litter on beaches of Amchitka Island, Alaska. Mar. Environ. Res. 3, 171–184. Merrell Jr., T.R., 1984. A decade of change in nets and plastic litter from fisheries in Alaska. Mar. Pollut. Bull. 15, 378–384. OSPAR Commission, 2007. OSPAR Pilot Project on Monitoring Marine Beach Litter: Monitoring of Marine Litter on Beaches in the OSPAR Region. OSPAR Commission, London. Ribic, C.A., Sheavly, S.B., Rugg, D.J., Erdmann, E.S., 2010. Trends and drivers of marine debris on the Atlantic coast of the United States 1997–2007. Mar. Pollut. Bull. 60, 1231–1242.
Ribic, C.A., Sheavly, S.B., Rugg, D.J., Erdmann, E.S., 2012. Trends in marine debris along the U.S. Pacific Coast and Hawai’i 1998–2007. Mar. Pollut. Bull. 64, 994– 1004. Ryan, P.G., Moloney, C.L., 1990. Plastic and other artefacts on South African beaches: temporal trends in abundance and composition. S. Afr. J. Sci. 86, 450–452. Ryan, P.G., Moore, C.J., van Franeker, J.A., Moloney, C.L., 2009. Monitoring the abundance of plastic debris in the marine environment. Phil. Trans. R. Soc. B 364, 1999–2012. Sheavly, S.B., 2007. National Marine Debris Monitoring Program: Final Program Report, Data Analysis and Summary. Ocean Conservancy, Washington, DC. Slip, D.J., Burton, H.R., 1991. Accumulation of fishing debris, plastic litter, and other artefacts, on Heard and Macquarie Islands in the Southern Ocean. Environ. Conserv. 18, 249–254. Swanepoel, D., 1995. An analysis of beach debris accumulation in Table Bay, Cape Town, South Africa. MSc Thesis, University of Cape Town, Cape Town. Thornton, L., Jackson, N.L., 1998. Spatial and temporal variations in debris accumulation and composition on an estuarine shoreline, Cliffwood Beach, New Jersey, USA. Mar. Pollut. Bull. 36, 705–711. Walker, T.R., Reid, K., Arnould, J.P.Y., Croxall, J.P., 1997. Marine debris surveys at Bird Island, South Georgia 1990–1995. Mar. Pollut. Bull. 34, 61–65. Williams, A.T., Tudor, D.T., 2001. Litter burial and exhumation: spatial and temporal distribution on a cobble pocket beach. Mar. Pollut. Bull. 42, 1031–1039.
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
The effect of fine-scale sampling frequency on estimates of beach litter accumulation Peter G. Ryan a,⇑, Annerie Lamprecht b, Debbie Swanepoel a, Coleen L. Moloney b a b
Percy FitzPatrick Institute of African Ornithology, DST/NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa Marine Research Institute and Department of Biological Sciences, University of Cape Town, Rondebosch 7701, South Africa
a r t i c l e
i n f o
Article history: Available online 16 September 2014 Keywords: Stranded debris Plastics Sample rate Accumulation Meta-analyses South Africa
a b s t r a c t The effect of sampling frequency on estimates of the rate of litter accumulation was determined for two South African sandy beaches. After initial cleaning, all manufactured items >10 mm diameter were collected in alternating bouts of daily or weekly cleanups. Daily sampling collected 2.5 (range 2.1–3.4) times more litter items than weekly samples and 1.7 (1.3–2.3) times more litter by mass. Low density items such as foamed polystyrene showed a greater differential (4–5 times more items from daily sampling), presumably due to faster turnover of lightweight litter items. Variation in weekly samples was not consistently less than daily estimates, suggesting that less frequent samples only partly integrate short-term fluctuations in litter dynamics. Researchers using beach accumulation data to infer trends in nearshore marine litter, or to assess the efficacy of litter mitigation measures, need to ensure consistency in sampling frequency. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Plastic and other drifting artefacts are major marine pollutants, and considerable efforts have been made to prevent the release of plastic articles into the environment (e.g. Coe and Rogers, 1997). Assessing the efficacy of measures to reduce marine debris at sea is complicated by its great spatial heterogeneity and the high cost of at-sea surveys (Ryan et al., 2009). As a result, much of what we know about marine litter is inferred from surveys of litter stranding on beaches (e.g. Derraik, 2002; Barnes et al., 2009), and major regional initiatives to assess trends in marine debris rely heavily on beach monitoring (Sheavly, 2007; OSPAR Commission, 2007). Initial studies of beach litter reported standing stocks of accumulated litter, which can show gross changes in the abundance and distribution of litter (e.g. Ryan and Moloney, 1990), but are not necessarily a good indicator of the abundance of litter in adjacent coastal waters (Escardó-Boomsma et al., 1995; Ryan et al., 2009). Other factors that affect the amount of litter on beaches include the physical environment (beach structure, wave action and local currents, which influence both stranding and removal rates); beach dynamics (burial or exposure of litter); proximity to urban centres (which influences exposure to land-based sources
⇑ Corresponding author. Tel.: +27 21 6502966; fax: +27 21 6503295. E-mail address: [email protected] (P.G. Ryan). http://dx.doi.org/10.1016/j.marpolbul.2014.08.036 0025-326X/Ó 2014 Elsevier Ltd. All rights reserved.
such as the number of beach visitors and urban run-off); and formal or informal cleanup efforts (Merrell, 1980; Bowman et al., 1998; Williams and Tudor, 2001). Also, increases in the standing stock of beach litter may reflect long-term accumulation rather than a change in the amount of debris at sea (Ryan et al., 2009). As a result, estimates of litter accumulation rates made by repeatedly cleaning the same stretch of beach currently are best practice for using beach surveys to assess marine litter trends (e.g. Sheavly, 2007; OSPAR Commission, 2007; Ryan et al., 2009; Ribic et al., 2010, 2012). One problem with litter accumulation estimates is that they might be sensitive to the frequency with which litter is sampled because of differences in litter turn-over rates (Ryan et al., 2009). Dixon and Cooke (1977) showed that the weekly retention rate of marked bottles on beaches in Kent, UK, varied among beaches and was influenced by the type of bottle, because plastic bottles remained ashore longer than glass bottles. Subsequent studies have confirmed different retention rates depending on beach structure and the type of litter (Merrell, 1980; Garrity and Levings, 1993; Bowman et al., 1998; Kataoka et al., 2013). EscardóBoomsma et al. (1995) differentiated the amount of litter arriving on a beach (the loading rate) from the amount of litter that accumulates per unit of time (net accumulation rate). However, this distinction has been largely ignored by subsequent studies, which have reported accumulation rates at sampling frequencies ranging from daily (Eriksson et al., 2013), every two weeks (Williams and
250
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Tudor, 2001), monthly (Corbin and Singh, 1993; Thornton and Jackson, 1998; Sheavly, 2007), quarterly (OSPAR Commission, 2007), every six months (Slip and Burton, 1991) to annually (Merrell, 1984; Edyvane et al., 2004). Some studies have sampled at different frequencies in different seasons (e.g. Walker et al., 1997; Ivar do Sul et al., 2011). Sheavly (2007) reported no difference in accumulation rate for samples collected at different intervals, but Ribic et al. (2010, 2012) found that frequent sampling (614 days) and very infrequent sampling (>180 days) affected debris counts relative to their 28-day standard sampling interval. Ribic et al. (2010, 2012) failed to report how these changes in sampling frequency affected litter counts, but the inference was that too frequent sampling reduced litter counts, and very infrequent samples increased litter counts. However, Ryan et al. (2009) included results from an unpublished MSc dissertation (Swanepoel, 1995) to suggest that summing daily accumulation samples to obtain weekly data yielded up to sixtimes more litter items than weekly sampling at the same beaches. More recently, Eriksson et al. (2013) showed that daily sampling at sub-Antarctic Macquarie Island increased litter accumulation rates almost ten-fold compared to monthly sampling, although surveys took place in different years. These findings suggest that sampling frequency, at least at fine temporal scales, has a major impact on estimates of litter accumulation rates. In this paper we report the findings from Swanepoel’s (1995) work in more detail, and supplement them with additional sampling. We compare the amount (number and mass) of litter collected daily with weekly accumulations collected at the same sites to evaluate the extent to which more frequent sampling increases estimates of the amount of litter arriving on beaches. We also assess whether less frequent sampling reduces the variability in beach litter accumulation data by integrating fine-scale variations (given that processes such as weather, wave action and tidal patterns might be expected to affect litter arrival and exhumation rates at fine temporal scales). Our findings have relevance for the design of beach litter monitoring programmes, and are especially important for meta-analyses of beach litter accumulation data collected at different sampling frequencies.
2. Materials and methods The study took place at two beaches in Table Bay, South Africa (Fig. 1), in the Cape Town metropolitan area. The beaches were chosen to have similar physical structures but to be different distances from the urban centre and to have contrasting ease of access and user profiles. Milnerton Beach is a popular recreational beach 10 km from the city centre, used for walking, surfing, swimming and other recreational activities, and attracts large numbers of visitors daily, especially in summer (November–March). Koeberg Nature Reserve is 25 km farther north (35 km from the city centre) and the study beach is closed to visitors. Both are exposed, west-facing, fine-grained sandy beaches with gentle to moderate beach profiles (10–15° slopes), and both are backed by steep, vegetated dunes which limit shoreward migration of litter. Neither study site was cleaned by local authorities during the surveys, but adjacent beaches at both sites were cleaned: Milnerton was cleaned by municipal workers weekly (1994 and 1995) or daily (2012), whereas Koeberg was cleaned by volunteers every four months. We could not control informal cleaning efforts by members of the public. At each beach, a study area was demarcated, divided into ten equal subsections along the beach (Fig. 1). In 1994 and 1995, 500 m of beach was sampled (ten 50-m subsections) at each beach, and the same site was sampled at Koeberg in 2012. However, at Milnerton only the southern 250 m surveyed in the 1990s was sampled in 2012 (ten 25-m subsections) because the marked
increase in the density of litter prevented both beaches being sampled in full on the same day. Prior to each experiment, a team of volunteers removed all accumulated macro litter (articles >10 mm diameter) between the water line and the dune vegetation (including any litter visible in the vegetation) from each study area, and from adjacent buffer zones approximately 25 m wide at each end of the study sections. The following day (24 h, or two tidal cycles later), sampling of ‘new’ litter commenced at both sites. Teams of three to six observers worked systematically along each beach, collecting all litter items in each of the ten subsections. Each subsection was searched until no items had been found for several minutes, with searches of each section taking roughly 5–15 min, depending on the amount of litter present. In 1994, daily sampling continued for 15 days (18 October–1 November), and was followed by one sample after 4 days, one after 6 days and four weekly samples, and then another 14 daily samples (9–23 December). In 1995, the initial cleaning took place on 26 June, with a weekly sample on 3 July, 14 daily samples to 17 July, and then a final weekly sample on 24 July. In 2012, daily sampling continued for 10 days (2–11 October), followed by three weekly samples (18 and 25 October and 1 November), 10-days of daily sampling (2–11 November), four weekly samples (18 and 25 November and 2 and 9 December), and then a final 10 daily samples (10–19 December). There was no consistent difference in tidal phase (neap or spring) between daily and weekly sample periods. All litter samples were returned to the University of Cape Town to be sorted and counted. Each debris item was identified as far as possible and categorised by type of material (plastic, metal, glass, cigarette butts, etc.). Wood items were included if they were ‘worked’ rather than natural wood. Plastic items were further subdivided by function: packaging and other single-use items, user items including fishing gear, and plastic fragments of uncertain provenance. Foamed polystyrene items (including cups and fast-food trays as well as packing chips and moulded packaging) were placed in a separate category because of their lower density than most other litter items. Articles were dried and cleaned of sand before weighing. In 1994, the entire sample was weighed to the nearest 25 g with a Pesola spring balance, but large pieces of timber and other very large litter items were not weighed. No mass data were recorded in 1995, but in 2012 each litter category was weighed separately to the nearest 0.1 g on an electronic top-pan balance. Litter samples were kept separate for each 50- or 25-m subsection of beach and data compared among subsections to check for consistent differences in fine-scale accumulation of litter. Such differences were detected, presumably linked to local currents, wave action and resultant beach structure (Bowman et al., 1998), but there was no tendency for the end subsections at each beach to have the highest accumulation rates, suggesting that lateral drift from adjacent areas outside the sampling area was minimal (data not shown). Thereafter, samples were pooled for each beach and accumulation data were summarised as the number or mass of litter items recorded per metre of beach per day, calculated as the litter yield divided by the length of beach sampled (m) and by the number of days since the last collection (days). For comparative purposes, the rate estimated from the 6-day accumulation sample in 1994 was combined with the weekly samples, but the 4-day accumulation sample was discarded. The effects of daily versus weekly sampling on the accumulation of litter on the beaches was tested with a general linear model (GLM), with beach and year as covariates: 1 Accumulated litter ðm1 d Þ ¼ eb0 þ bweekly=daily þ bbeach þ byear þ e
where b0 is a constant; bweekly/daily, bbeach and byear are parameter sets describing the effects of the predictor variables (sampling
251
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
10 km Koeberg
Blouberg
Robben Island
Table View
Table Bay Rietvlei
Milnerton
Black River Cape Town
Fig. 1. Table Bay, South Africa, showing the location of the two study sites at Milnerton and Koeberg, each comprising 10 subsections.
frequency, beach and year) on litter accumulation, and e refers to the residuals (there were no significant interaction effects). The model was applied to all litter items and also to each major category of litter. Where there were zero values in the data set (for nonplastic items), a small constant (equal to 10% of the mean abundance) was added to each data point before a logarithmic transformation was applied to reduce heteroscedasticity and normalise the data (both checked by residual analysis). Coefficients of variation corrected for sample size [=SD/mean ⁄ (1 + (1/4n))] were used to compare relative variability in the number and mass of litter items recorded in daily and weekly samples from the same beach/ year. However, there were too few weekly samples in July 1995 to compare their variation with daily samples.
3. Results In October–December 1994, the 34 accumulation samples at both beaches comprised 38,155 litter items weighing at least 151.1 kg. In July the following year 39,963 items were sampled in only 16 collections, and 37 collections in 2012 yielded 199,303 items weighing 257.8 kg. Plastic articles dominated the samples
numerically from both beaches in all three years, comprising 86.5% by number in 1994, 89.3% in 1995 and 93.5% in 2012. The contribution of plastics to the mass of litter was only recorded in 2012, when they comprised 58.9% overall, with wood (21.9% of total mass) and glass (12.8%) dominating the mass of non-plastic items. Plastics were more dominant at Koeberg (95.3% by number in all years and 62.0% by mass in 2012), which is farther from the city centre and has many fewer visitors than Milnerton (91.3% by number and 57.7% by mass). After plastics, wood contributed most to the mass of litter at both Koeberg (30.4%) and Milnerton (18.4%), but glass items were more important at Milnerton (16.9% of total litter mass) than at Koeberg (2.7%). The GLM indicated that estimates of daily accumulation rates of all litter items varied with sample site (F = 335, df = 1, 169, p < 0.00001), year (F = 81, df = 2, 169, p < 0.00001) and sampling frequency, with daily sampling providing consistently greater estimates than weekly sampling (F = 32, df = 1, 169, p < 0.00001; Table 1). On average, after accounting for the effects of beach and year, daily samples captured 2.5 times more litter overall than weekly samples (Fig. 2). This factor was consistent (1.8–2.5) for all broad litter categories apart from foamed polystyrene (4.2), which had the largest factor of all litter types in five of six beach-year
252
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Table 1 The effect of sampling frequency (daily versus weekly) on estimates of litter accumulation rates (number of items per metre of beach per day) at two South African beaches in three years. ‘Factor’ is the number of times the rate estimated from daily samples exceeds that from weekly samples (=daily rate/weekly rate). Site and litter type
1994
1995
2012
Day n = 29
Week n=5
Factor
Day n = 14
Week n=2
Factor
Day n = 30
Week n=7
Factor
0.40 0.71 0.15 0.05 1.31
0.22 0.11 0.04 0.01 0.39
1.8 6.7 3.3 4.2 3.4
1.35 0.37 0.47 0.27 2.46
0.60 0.04 0.13 0.10 0.86
2.2 10.5 3.7 2.7 2.8
4.32 1.38 1.54 6.26 13.50
1.63 0.32 0.89 2.27 5.11
2.6 4.3 1.7 2.8 2.6
Cigarettes Wood Other non-plastics All non-plastics
0.16 0.04 0.04 0.24
0.05 0.01 0.01 0.07
3.4 4.0 3.1 3.5
0.18 0.06 0.17 0.41
0.02 0.01 0.03 0.06
7.6 8.3 5.6 6.7
0.57 0.13 0.38 1.08
0.21 0.05 0.09 0.36
2.7 2.5 4.1 3.0
All litter
1.55
0.46
3.4
2.87
0.93
3.1
14.58
5.47
2.7
0.11 0.16 0.04 0.02 0.33
0.05 0.07 0.02 0.01 0.15
2.1 2.1 2.3 2.2 2.1
0.38 0.50 0.21 0.12 1.21
0.26 0.10 0.15 0.06 0.56
1.4 5.1 1.5 2.1 2.1
0.75 0.50 0.27 0.48 1.99
0.32 0.09 0.10 0.10 0.61
2.3 5.7 2.7 4.6 3.3
Cigarettes Wood Other non-plastics All non-plastics
0.01 0.01 0.01 0.03
0.00 0.01 0.00 0.01
3.7 2.0 4.9 2.8
0.01 0.04 0.04 0.09
0.00 0.02 0.03 0.06
2.9 1.9 1.2 1.5
0.01 0.02 0.02 0.05
0.00 0.01 0.01 0.02
3.0 2.5 3.3 2.8
All litter
0.35
0.16
2.2
1.30
0.62
2.1
2.04
0.63
3.2
Milnerton Packaging Polystyrene User items Fragments All plastics
Koeberg Packaging Polystyrene User items Fragments All plastics
comparisons (Table 1). The consistent pattern of crude factors when comparing daily to weekly estimates for each site and year (Table 1) provides further evidence for the generality of this effect. Sampling frequency appears to have slightly less of an impact on estimates of the mass of litter accumulating on beaches. The crude mass estimates in 1994 indicated that daily masses at both beaches were 1.5 times greater than estimates based on weekly samples. These values increased to 1.8 (Koeberg) and 2.3 (Milnerton) in 2012, when the average mass of each litter item (1.3 g) was considerably less than in 1994 (4.0 g), despite the fact that large pieces of wood and other very large litter items were not weighed in 1994. Data on how the masses of different categories of litter differed in their response to sampling frequency were restricted to 2012. Plastic litter items tended to have a slightly greater day:week factor (Milnerton 2.5, Koeberg 1.9) than non-plastic items (Milnerton 2.1, Koeberg 1.7). Foamed polystyrene again had the greatest factors (Milnerton 3.2, Koeberg 3.3), although these were lower than the factors for numbers of polystyrene items in 2012 (Table 1). The numbers of litter items accumulating daily in different time periods at each beach varied considerably (Table 2), with an average coefficient of variation of 59%. This was slightly greater than the average weekly coefficients of variation in 1994 and 2012 (49%), but there was no consistent pattern comparing daily and weekly CVs. Daily variation in the mass of litter (64%) also was slightly greater than weekly samples (49%), but again there was no consistent pattern. For example, in 1994, the daily CV in mass was greater than the weekly CV at Koeberg, but weekly variability was greater at Milnerton (Table 2). Coefficients of variation tended to be greater at Koeberg than at Milnerton (Table 2), possibly linked to the smaller volumes of litter at Koeberg. In 1994/95, daily CVs were greatest in winter (July) and decreased through spring (October) to summer (December). However, in 2012 CVs of daily litter counts increased from October to December (but there was no corresponding pattern in CVs of litter masses; Table 2).
4. Discussion Beaches are dynamic systems where litter can be stranded by the sea; dumped by beach visitors; buried and exhumed; exported inland, offshore or along-shore by wind, wave or tidal action; or removed by formal or informal beach cleaning efforts (Ryan et al., 2009). The best way to monitor the persistence of litter items (and, by inference, estimate the scale of these movements and the resultant litter turnover rates) is by mark-recapture studies (e.g. Merrell, 1980; Garrity and Levings, 1993; Bowman et al., 1998; Williams and Tudor, 2001; Kataoka et al., 2013). However, markrecapture studies are logistically challenging on a large scale at urban beaches characterised by large litter loads; most studies to date have only sampled specific litter types (Dixon and Cooke, 1977; Merrell, 1980; Williams and Tudor, 2001). Our study sampled over 280,000 litter items; attempting to mark and follow the fate of even a small subsample of this number would be almost impossible. By alternating daily and weekly beach litter collections we were able to show that more frequent sampling increased the estimate of the amount of litter found on a beach. Our results were consistent across beaches in three different years, sampling in both summer and winter, with little variation in the magnitude of the effect of more frequent sampling. For most litter items, daily sampling resulted in 2–3 times more litter being collected than weekly sampling. This presumably results from rapid turnover of many litter items which are missed in weekly samples because they are either buried or transported off the beach (blown inshore, blown or washed offshore, carried alongshore, or removed by informal beach cleaning efforts by wellmeaning members of the public). Small, lightweight items are more likely to be blown inland or buried, resulting in rapid turnover ( Merrell, 1984; Williams and Tudor, 2001), and hence should have a greater differential between daily and weekly estimates of accumulation rates. This was shown by foamed polystyrene, which is particularly prone to wind dispersal and had more than a fourfold increase in litter accumulation rates based on daily samples.
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Plastics
In addition to differences among litter types, the effect of sampling frequency on estimates of litter arrival rates varies with the magnitude of the difference in sampling frequency. Thus Eriksson et al. (2013) found almost an order of magnitude more litter by sampling daily rather than monthly at beaches on two sub-Antarctic islands. The size of the effect probably also varies between beaches, as a result of differences in turnover rates linked to beach structure (e.g. slope, orientation, exposure) and local conditions (e.g. wind strength and direction). In general, the faster the turnover rate, the greater the disparity between arrival rates estimated at different sampling frequencies (cf. Kataoka et al., 2013). The effect of daily sampling was greater in 2012 than in the 1990s, presumably because the mean mass of litter items in 2012 was lower, increasing turnover rates. Our findings are particularly important for meta-analyses comparing beach litter accumulation data collected at different frequencies. For example, Barnes and Milner (2005) presented broad latitudinal patterns in beach litter density, ostensibly representing annual accumulation rates, but scrutiny of their sources indicate a diversity of sampling approaches from standing stock surveys to seasonal and monthly sampling. Even if analyses control for sampling frequencies, the intrinsic differences in litter turnover rates among beaches (see above) are likely to confound comparisons unless samples are collected at a high frequency (e.g. daily). Unfortunately, daily sampling is costly to implement because it is time and labour intensive. Less frequent, regular sampling (e.g. monthly) of the same site or set of sites is adequate to track long-term changes in litter loads. Such programmes would benefit from short bouts of daily sampling to assess the extent to which infrequent sampling underestimates the loading rates of specific litter items (i.e. to evaluate the proportions of various litter items arriving on a beach that actually accumulate there). This information could be used to develop correction factors to improve the accuracy of monitoring programmes, although turnover rates probably depend on the size as well as type of litter item. Ideally, litter comparisons among beaches should be based on estimates of loading rate rather than accumulation rate.
Non-plastics 0.10
0.8 0.08 0.6 0.06 0.4
Litter accumulation rate (items per metre per day)
0.04 0.2
0.02
0.0
0.00
Packaging
Polystyrene 0.25
0.3 0.20 0.15
0.2
0.10 0.1 0.05 0.00
0.0
Fragments
User items
0.05 0.10 0.04 0.08 0.03 0.06 0.02
0.04
0.01
0.02
Acknowledgements
0.00
0.00
Day
Week
Day
Week
Fig. 2. The effect of sampling frequency (daily versus weekly) on estimates of litter accumulation rates for different categories of beach litter determined by a GLM including data from two Cape Town beaches over three years. Error bars show 95% confidence intervals.
Table 2 Coefficients of variation (CV) in daily and weekly litter collections at two South African beaches in three years. Too few weekly samples were collected in July 1995 to estimate the weekly CV, and there were no mass data for 1995. Year and month
Milnerton
253
Koeberg
Number
Mass
Number
1994 October daily (n = 15) December daily (n = 14) Weekly (n = 5)
48% 24% 27%
59% 44% 71%
1995 July daily (n = 14)
89%
–
130%
–
2012 October daily (n = 10) November daily (n = 10) December daily (n = 10) Weekly (n = 7)
38% 49% 61% 44%
67% 27% 23% 23%
36% 40% 102% 59%
94% 77% 88% 54%
47% 42% 65%
Mass 87% 71% 57%
We thank the numerous volunteers who assisted with litter collections, especially P. Hardcastle, M. Lamprecht, A. Legodi, T. Oosthuizen and J. Oosthuizen. G. Greef, J. Le Roux and H. Westman kindly gave permission to work in the Koeberg Private Nature Reserve. G. Frost (Milnerton Municipality) and J. Kieser (Plastics SA) supplied rubbish bags and assisted with initial cleanups. The National Research Foundation – South Africa and the University of Cape Town – South Africa provided funds for field expenses and bursary support for A.L. and D.S.
References Barnes, D.K.A., Milner, P., 2005. Drifting plastic and its consequences for sessile organism dispersal in the Atlantic Ocean. Mar. Biol. 146, 815–825. Barnes, D.K.A., Galgani, F., Thompson, R.C., Barlaz, M., 2009. Accumulation and fragmentation of plastic debris in global environments. Phil. Trans. R. Soc. B 364, 1985–1998. Bowman, D., Manor-Samsonov, N., Golit, A., 1998. Dynamics of litter pollution on Israeli Mediterranean beaches: a budgetary, litter flux approach. J. Coastal Res. 14, 2588–2597. Coe, J.M., Rogers, D.B., 1997. Marine Debris: Sources, Impacts and Solutions. Springer-Verlag, New York. Corbin, C.J., Singh, J.G., 1993. Marine debris contamination of beaches in St. Lucia and Dominica. Mar. Pollut. Bull. 26, 325–328. Derraik, J.G.B., 2002. The pollution of the marine environment by plastic debris: a review. Mar. Pollut. Bull. 44, 842–852. Dixon, T.R., Cooke, A.J., 1977. Discarded containers on a Kent Beach. Mar. Pollut. Bull. 8, 105–109.
254
P.G. Ryan et al. / Marine Pollution Bulletin 88 (2014) 249–254
Edyvane, K.S., Dalgetty, A., Hone, P.W., Higham, J.S., Wace, N.M., 2004. Long-term marine litter monitoring in the remote Great Australian Bight, South Australia. Mar. Pollut. Bull. 48, 1060–1075. Eriksson, C., Burton, H., Fitch, S., Schulz, M., 2013. Daily accumulation rates of marine debris on sub-Antarctic island beaches. Mar. Pollut. Bull. 66, 199–208. Escardó-Boomsma, J., O’Hara, K., Ribic, C.A., 1995. National Marine Debris Monitoring Program, vols. 1–2. US EPA Office of Water, Washington, DC. Garrity, S.D., Levings, S.C., 1993. Marine debris along the Caribbean coast of Panama. Mar. Pollut. Bull. 26, 317–324. Ivar do Sul, J.A., Barnes, D.K.A., Costa, M.F., Convey, P., Costa, E.S., Campos, L., 2011. Plastics in the Antarctic environment: are we looking only at the tip of the iceberg? Oecol. Aust. 15, 150–170. Kataoka, T., Hinata, H., Kato, S., 2013. Analysis of a beach as a time-invariant linear input/output system of marine litter. Mar. Pollut. Bull. 77, 266–273. Merrell Jr., T.R., 1980. Accumulation of plastic litter on beaches of Amchitka Island, Alaska. Mar. Environ. Res. 3, 171–184. Merrell Jr., T.R., 1984. A decade of change in nets and plastic litter from fisheries in Alaska. Mar. Pollut. Bull. 15, 378–384. OSPAR Commission, 2007. OSPAR Pilot Project on Monitoring Marine Beach Litter: Monitoring of Marine Litter on Beaches in the OSPAR Region. OSPAR Commission, London. Ribic, C.A., Sheavly, S.B., Rugg, D.J., Erdmann, E.S., 2010. Trends and drivers of marine debris on the Atlantic coast of the United States 1997–2007. Mar. Pollut. Bull. 60, 1231–1242.
Ribic, C.A., Sheavly, S.B., Rugg, D.J., Erdmann, E.S., 2012. Trends in marine debris along the U.S. Pacific Coast and Hawai’i 1998–2007. Mar. Pollut. Bull. 64, 994– 1004. Ryan, P.G., Moloney, C.L., 1990. Plastic and other artefacts on South African beaches: temporal trends in abundance and composition. S. Afr. J. Sci. 86, 450–452. Ryan, P.G., Moore, C.J., van Franeker, J.A., Moloney, C.L., 2009. Monitoring the abundance of plastic debris in the marine environment. Phil. Trans. R. Soc. B 364, 1999–2012. Sheavly, S.B., 2007. National Marine Debris Monitoring Program: Final Program Report, Data Analysis and Summary. Ocean Conservancy, Washington, DC. Slip, D.J., Burton, H.R., 1991. Accumulation of fishing debris, plastic litter, and other artefacts, on Heard and Macquarie Islands in the Southern Ocean. Environ. Conserv. 18, 249–254. Swanepoel, D., 1995. An analysis of beach debris accumulation in Table Bay, Cape Town, South Africa. MSc Thesis, University of Cape Town, Cape Town. Thornton, L., Jackson, N.L., 1998. Spatial and temporal variations in debris accumulation and composition on an estuarine shoreline, Cliffwood Beach, New Jersey, USA. Mar. Pollut. Bull. 36, 705–711. Walker, T.R., Reid, K., Arnould, J.P.Y., Croxall, J.P., 1997. Marine debris surveys at Bird Island, South Georgia 1990–1995. Mar. Pollut. Bull. 34, 61–65. Williams, A.T., Tudor, D.T., 2001. Litter burial and exhumation: spatial and temporal distribution on a cobble pocket beach. Mar. Pollut. Bull. 42, 1031–1039.
