181
sampling, e.g., of a section of a beach, including the whole
vertical and horizontal dimension, could be an option, although
not yet conducted for MPs. In any case, care should be taken
when formulating research questions, as this will set the
framework for considerations regarding the sampling design.
Spatial and Temporal Replication
To get a representative sample, care needs to be taken with
respect to appropriate replication as well as the amount of
sample, which will be taken. If study areas of various sizes
are compared, it needs to be considered, whether the number
of replicates is kept the same or whether they are adjusted to
the area (balanced vs. unbalanced sampling design). For
beach sediment, Kim et al. (2015) adjusted sampling effort
to beach size, whereas the majority of studies kept replicate
numbers the same. In the current literature, replicate samples
for one beach can range from one to 88 (Besley et al. 2017),
whereas recommendations suggest a replication of at least
five (Hanke et al. 2013). For beach sediments, Dekiff et al.
(2014) found no significant variability in MP abundance
within a 100 m transect, taking six replicate samples. Low
spatial variability on a small scale (within tens of m) was
further found in a recent study from Fisner et al. (2017) on
plastic pellets (~ 1–6 mm; (Hidalgo-Ruz et al. 2012), whereas
this study further found a high spatial variability on a large
scale (within km). Contrary, Besley et al. (2017), including
smaller MPs (300–5000 μm), found a high spatial variability
among ten samples on a transect of 100 m. Confidence intervals around the mean in this study decreased rapidly after a
replication of five, and 11 replicates would be needed to
reach a 0.5 standard deviation at a confidence level of 90%
(Besley et al. 2017). Those results are supported by a further
study concentrating on large MPs (1–5 mm) on a 100 m transect on a tropical beach (six replicates; (Imhof et al. 2017).
For surface water samples there is one study investigating
spatial variability within the eastern North Pacific, off
California (~ 20°–40°N, 120°–155°W; (Goldstein et al.
2013). They found that MP concentrations were highly variable over relatively small scales (tens of km) as well as for
large scales (hundreds to thousands of km).
It is also stated that MP abundance varies over numerous
temporal scales and detection of temporal trends are often
hampered by the sampling design (Browne et al. 2015).
Recent studies conducted on beaches found high daily variability due to tidal dynamics (Moreira et al. 2016; Imhof
et al. 2017). One possibility to improve knowledge about
temporal patterns could be through ice or sediment cores
(Costa and Duarte 2017), by analyzing different layers separately. For the water surface, high inter-annual variability
was found (Law et al. 2010; Doyle et al. 2011; Law et al.
2014), whereas Law et al. (2010), investigating a 22-year
dataset of surface plankton net tows, found no strong temporal trends in MP concentrations within this data set.
Nevertheless, the time span needed for a sampling campaign
should be considered beforehand. For example for beach
sediment sampling, sampling periods range over several
hours to years (Browne et al. 2015).Whereas for some study
questions, sampling over a certain period of time may not be
a problem, for others it could lead to biased results. This
might, for instance, apply to the sampling of various river
mouths at a delta over several days. Strongly changing precipitation between sampling days could hamper comparability, as MP runoff could be enhanced during days of heavy
rainfall, similar to what was hypothesized in a recent study
comparing MP load of waste water treatment plants effluents
on two different dates with differing participation events
(Primpke et al. 2017a).
Sampling Depth
For both, sediments and water column, the optimal sampling
depth remains another open question. Sediment sampling is
recommended to a depth of at least 5 cm (Hanke et al. 2013;
Besley et al. 2017), whereas studies report that a potential
proportion can be lost if deeper sediment layers are not sampled (Carson et al. 2011; Claessens et al. 2011). Thus, it has
already been stated that samples should be taken at a depth to
1 m, to get a more precise picture of MP abundances (Turra
et al. 2014; Fisner et al. 2017). For the water column, only
few studies exist where different depths were concurrently
sampled (Lattin et al. 2004; Reisser et al. 2015). In one study,
no significant differences were found between the sea surface, the water column (5 m depth), and above the bottom
(Lattin et al. 2004), whereas the other found that MP concentrations decreased exponentially, with highest amounts
within the first 0.5 m of the water column (Reisser et al.
(2015). This is confirmed by Goldstein et al. (2013), detecting the highest concentrations of MPs during low wind conditions, when minimal mixing occurs between shallow and
deeper water layers. The optimal sampling depth will finally
be a compromise between increasing sampling surface and
sampling depth and thus will also be determined by the
research question.
Reporting of Data
Though different methods are necessary depending on the
research question, researchers should aim for standardization, the most important one being size classes and reporting
units. Regarding size classes the upper limit for MPs is
5 mm, whereas the lower limit will be defined by the
sampling device, as well as the analytical method. Initial
Microplastics in Aquatic Systems – Monitoring Methods and Biological Consequences
sampling, e.g., of a section of a beach, including the whole
vertical and horizontal dimension, could be an option, although
not yet conducted for MPs. In any case, care should be taken
when formulating research questions, as this will set the
framework for considerations regarding the sampling design.
Spatial and Temporal Replication
To get a representative sample, care needs to be taken with
respect to appropriate replication as well as the amount of
sample, which will be taken. If study areas of various sizes
are compared, it needs to be considered, whether the number
of replicates is kept the same or whether they are adjusted to
the area (balanced vs. unbalanced sampling design). For
beach sediment, Kim et al. (2015) adjusted sampling effort
to beach size, whereas the majority of studies kept replicate
numbers the same. In the current literature, replicate samples
for one beach can range from one to 88 (Besley et al. 2017),
whereas recommendations suggest a replication of at least
five (Hanke et al. 2013). For beach sediments, Dekiff et al.
(2014) found no significant variability in MP abundance
within a 100 m transect, taking six replicate samples. Low
spatial variability on a small scale (within tens of m) was
further found in a recent study from Fisner et al. (2017) on
plastic pellets (~ 1–6 mm; (Hidalgo-Ruz et al. 2012), whereas
this study further found a high spatial variability on a large
scale (within km). Contrary, Besley et al. (2017), including
smaller MPs (300–5000 μm), found a high spatial variability
among ten samples on a transect of 100 m. Confidence intervals around the mean in this study decreased rapidly after a
replication of five, and 11 replicates would be needed to
reach a 0.5 standard deviation at a confidence level of 90%
(Besley et al. 2017). Those results are supported by a further
study concentrating on large MPs (1–5 mm) on a 100 m transect on a tropical beach (six replicates; (Imhof et al. 2017).
For surface water samples there is one study investigating
spatial variability within the eastern North Pacific, off
California (~ 20°–40°N, 120°–155°W; (Goldstein et al.
2013). They found that MP concentrations were highly variable over relatively small scales (tens of km) as well as for
large scales (hundreds to thousands of km).
It is also stated that MP abundance varies over numerous
temporal scales and detection of temporal trends are often
hampered by the sampling design (Browne et al. 2015).
Recent studies conducted on beaches found high daily variability due to tidal dynamics (Moreira et al. 2016; Imhof
et al. 2017). One possibility to improve knowledge about
temporal patterns could be through ice or sediment cores
(Costa and Duarte 2017), by analyzing different layers separately. For the water surface, high inter-annual variability
was found (Law et al. 2010; Doyle et al. 2011; Law et al.
2014), whereas Law et al. (2010), investigating a 22-year
dataset of surface plankton net tows, found no strong temporal trends in MP concentrations within this data set.
Nevertheless, the time span needed for a sampling campaign
should be considered beforehand. For example for beach
sediment sampling, sampling periods range over several
hours to years (Browne et al. 2015).Whereas for some study
questions, sampling over a certain period of time may not be
a problem, for others it could lead to biased results. This
might, for instance, apply to the sampling of various river
mouths at a delta over several days. Strongly changing precipitation between sampling days could hamper comparability, as MP runoff could be enhanced during days of heavy
rainfall, similar to what was hypothesized in a recent study
comparing MP load of waste water treatment plants effluents
on two different dates with differing participation events
(Primpke et al. 2017a).
Sampling Depth
For both, sediments and water column, the optimal sampling
depth remains another open question. Sediment sampling is
recommended to a depth of at least 5 cm (Hanke et al. 2013;
Besley et al. 2017), whereas studies report that a potential
proportion can be lost if deeper sediment layers are not sampled (Carson et al. 2011; Claessens et al. 2011). Thus, it has
already been stated that samples should be taken at a depth to
1 m, to get a more precise picture of MP abundances (Turra
et al. 2014; Fisner et al. 2017). For the water column, only
few studies exist where different depths were concurrently
sampled (Lattin et al. 2004; Reisser et al. 2015). In one study,
no significant differences were found between the sea surface, the water column (5 m depth), and above the bottom
(Lattin et al. 2004), whereas the other found that MP concentrations decreased exponentially, with highest amounts
within the first 0.5 m of the water column (Reisser et al.
(2015). This is confirmed by Goldstein et al. (2013), detecting the highest concentrations of MPs during low wind conditions, when minimal mixing occurs between shallow and
deeper water layers. The optimal sampling depth will finally
be a compromise between increasing sampling surface and
sampling depth and thus will also be determined by the
research question.
Reporting of Data
Though different methods are necessary depending on the
research question, researchers should aim for standardization, the most important one being size classes and reporting
units. Regarding size classes the upper limit for MPs is
5 mm, whereas the lower limit will be defined by the
sampling device, as well as the analytical method. Initial
Microplastics in Aquatic Systems – Monitoring Methods and Biological Consequences