Marine Pollution Bulletin 131 (2018) 212–217

Contents lists available at ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier.com/locate/marpolbul

Optimising beached litter monitoring protocols through aerial imagery T

a,⁎ a a b
A. Deidun , A. Gauci , S. Lagorio , F. Galgani
a
Department of Geosciences, Faculty of Science, University of Malta Tal-Qroqq campus, Msida MSD 2080, Malta
b
Station de Corse - IFREMER - Immeuble Agostini- Z.I. Furiani, 20600 Bastia, Corsica, France

A R T I C LE I N FO A B S T R A C T

Keywords: The monitoring of beached litter along the coast is an onerous obligation enshrined within a number of legis-
Drones lative frameworks (e.g. the MSFD) and which requires substantial human resources in the field. Through this
Monitoring obligations study, we have optimised the protocol for the monitoring of the same litter along coastal stretches within an
Marine litter MPA in the Maltese Islands through aerial drones, with the aim of generating density maps for the beached litter,
Optimised protocol
of assisting in the identification of the same litter and of mainstreaming this type of methodology within national
MSFD
and regional monitoring programmes for marine litter. Concurrent and concomitant in situ monitoring of
MPAs
beached litter enabled us to ground truth the aerial imagery results. Results were finally discussed within the
context of current and future MSFD monitoring obligations, with considerations made on possible future policy
implications.

1. Introduction information on litter quantities and distribution within the Mediterra-

Marine litter is considered nowadays as one of the most insidious
problems affecting the oceans, in particular since its accumulation has
been identified in all the main oceanic gyres (Suaria and Aliani, 2014).
Also, unfortunately, plastic pollution is globally distributed across all
oceans due to its properties of buoyancy and durability (Eriksen et al.,
2014). Since 1950, the global production of plastic has increased ra-
pidly and this upward trend in production will continue in the coming
decades (Cózar et al., 2014). The abundance of Marine Litter in the
Mediterranean Sea is increasing: in fact, it has been estimated that in
30 years, the abundance of floating Marine Litter will increase by 6–8%
when compared with the current situation (Lebreton et al., 2012).
Therefore, it is essential to monitor the Mediterranean Sea on an op-
erational basis in order to identify trends in marine litter accumulation.
Operationality can only be achieved through the adoption of innovative
monitoring means, and this study represents a preliminary attempt to
apply to selected coastal sites in the Maltese Islands an innovative
surface marine litter monitoring technique, involving drone flights.
Although no macroscale hydrodynamic phenomena leading to litter
accumulation, such as gyres, operate in the Mediterranean Sea, marine
litter is still considered of critical importance within the semi-enclosed
Basin due to the basin's limited exchanges with other oceans, its densely
populated coasts, highly developed tourism, 30% of the world's mar-
itime traffic passing through and various additional inputs of litter from
rivers and very urbanized areas (UNEP/MAP, 2015). Despite this, most
nean is still fragmented (Ioakeimidis et al., 2014).
The insidious nature of the marine occurrence of litter is acknowl-
edged through the inclusion of Marine Litter as one of the Marine
Strategy Framework Directive's (MSFD) 11 Descriptors. The list of in-
dicators approved by the EU Commission for the first six-year MSFD
cycle includes ‘Trends in the amount of litter washed ashore and/or
deposited on coastlines, including analysis of its composition, spatial
distribution and, where possible, source’ (Indicator 10.1.1). The MSFD
Initial Assessment submitted by the Maltese Islands for Descriptor 10
specifies that information on litter from the archipelago's coast emerges
mainly from one-off research initiatives or from clean-up efforts rather
than from regular or even operational monitoring initiatives, revealing
a glaring gap in the national monitoring capacity for this issue.
The approved MSFD monitoring programme for the Maltese Islands
proposes that, every three months, visual surveys by human observers
are conducted on two recreational beaches along two 100 m-long
transects located on each beach and that SCUBA diving surveys are
conducted along two transects within the nearshore area of two remote
beaches so as to collect data for indicator 10.1.1. Additionally, surveys
of floating litter (‘litter in the water column’), conducted from the same
beaches, are also envisaged. These monitoring obligations are onerous,
especially for a small island state, and necessitate considerable human
and infrastructure resources, also since the acquired data may not al-
ways be utilisable for additional marine litter reporting obligations that
the Maltese Islands have under the Integrated Monitoring and

⁎
Corresponding author at: Physical Oceanography Research Group, Department of Geosciences, Room 315, Chemistry Building 3rd Floor, University of Malta Tal-Qroqq campus, Msida
MSD 2080, Malta.
E-mail address: alan.deidun@um.edu.mt (A. Deidun).

https://doi.org/10.1016/j.marpolbul.2018.04.033
Received 15 February 2018; Received in revised form 10 April 2018; Accepted 12 April 2018
Available online 17 April 2018
0025-326X/ © 2018 Published by Elsevier Ltd.
A. Deidun et al.
Assessment Programme (ImAp) within the EcAp (Ecosystem Approach)
of the Barcelona Convention, under the Bathing Water Directive, for
Annex V of the International Convention for the Prevention of Pollution
from Ships (MARPOL) and for the Blue Flag programme.
Despite their relatively recent emergence, drones are increasingly
being deployed within research efforts. Research applications of drones
include the monitoring of beach topography (e.g. Casella et al., 2016),
the detection of oil spills, through integration with photogrammetry
(e.g. Capolupo et al., 2014) and the long-term monitoring of forests
(e.g. Zhang et al., 2016). Linchant et al. (2015) give an overview of the
various applications of Unmanned Aircraft Systems (UASs) in wildlife
monitoring. The drone deployment, however, only constitutes the in-
itial stage within an integrated process which involves extensive post-
processing of the raw data (e.g. aerial imagery, positioning data) pro-
vided by the drone. The regular acquisition of geo-referenced and high-
resolution aerial imagery by drones is both cost-effective and rapid,
assisting in the achievement of environmental monitoring obligations
associated with marine litter, especially when coupled with machine
learning protocols which increase the reliability of results by increasing
replicability and decreasing human bias.
The main goal of this work was the definition of a protocol for the
detection and monitoring of marine litter through the use of an off-the-
shelf aerial drone. Thus, the study's two main aims were the following:
(i) the formulation of an optimised protocol providing guidance on how
to spot beached and floating marine litter through the use of a drone
(unmanned aerial vehicle) and (ii) the application of machine learning
techniques (specifically, those related to image analyses) to the quali-
tative and quantitative analyses of aerial imagery for coastal litter
monitoring purposes. A set of recommendations on the integration of
protocols presented in this study within the monitoring indicators being
currently formulated for the second six-year MSFD cycle are also pro-
vided.
2. Materials and methods
2.1. Site description
Three coastal stretches situated along the coast of the North-East
Marine Protected Area of the Maltese Islands (Malta, Comino, Gozo)
were selected. These three low-lying rocky coastal stretches are located
at Bahar Ic-Caghaq, where the western and eastern flanks of a rocky
peninsula were monitored, and at Qawra Point, with the two locations
being considered, at least a prima facie, as coastal hot spots for the
beaching of litter. The domain was set such as to include marine as well
as terrestrial facets of the coastal zone, with the study sites being chosen
so as to represent (i) gently-/low-lying coasts congenial for the
beaching of litter, (ii) stretches of the coast which are not sheltered
from wind and wave exposure and (iii) easily-accessible sites which
enable repeated visits. This approach permits the joint identification of
anthropogenic litter on the coast, i.e. litter that has beached, as well as
floating debris within nearshore waters.
As a proof of concept, the protocol was tested on four sites around
the Maltese Islands, as shown in Fig. 1. Embayment regions within the
North-East Marine Protected Area that are frequently exposed to strong
winds and waves, were selected due to their ease of access and low-
lying topography (and thus high potential for beaching of litter) and
monitored over a period of three months. In particular, Qawra Point
was selected as this consists of north and south facing coastal stretches
that could be surveyed in parallel. On the other hand, Baħar iċ-Ċagħaq
was considered due to the occurrence of two rocky coastal stretches that
face the west and the east due to their position on opposing sides of a
headland. Fig. 2 shows the flight transections in both locations.
2.2. Drone-mediated monitoring
The methodology followed specifically targeted the requirements of
213
Marine Pollution Bulletin 131 (2018) 212–217
the Marine Strategy Framework Directive (MSFD) for the quantification
of beached and floating litter. Apart from the accumulation of marine
debris, the proposed technique allows an investigation of how this
changes with time. High resolution overlapping aerial images were
collected over fixed transects that cover coastal areas of about 200 m by
200 m (40 km2).
All collected data was mosaiced into one georeferenced orthophoto
map. Detected debris was then classified by position, material, and size.
The litter size categories adopted in this study are consistent with those
prescribed by the MSFD, namely the 2.5 cm to 5 cm, 5 cm to 10 cm,
10 cm to 20 cm, 20 cm to 30 cm and the 30 cm to 50 cm ones.
The aircraft used in this work was a DJI Phantom 4 Pro. The gimbal
was set to −90° to look at nadir. This allowed the capturing of photos
perpendicular to the direction of flight. Images with a resolution of
5472 × 3078 pixels were recorded. The shutter speed and aperture
were set to 1/320 s and f/2.8 respectively. The ISO setting was kept
fixed at 1000 so as to avoid changes in the intensity levels between
images. The drone was set to fly at an altitude of 30 m. Such a value was
empirically established as a compensation between image and spatial
resolution after carrying out tests at different altitudes ranging between
20 m and 50 m. While aerial imagery data collected from low altitudes
had a very good resolution, this only covered a very small area. Such a
scheme also required a very slow flight speed and a high photo capture
rate. Moreover, flights at low altitude attracted more attention from
beach users due to the propeller's operating noise. Images recorded
from a high altitude covered a larger area but suffered from a poor
resolution. As a result, the optimum altitude for data collection was
found to be 30 m. From this altitude, each frame covered an estimated
ground area of 45 m by 25 m. Automatically capturing images every 3 s
while the drone is flying at a speed of 3.5 m/s resulted in the frames
overlapping by 58% along the wider side and by 44% along the shorter
side. Data collection during the summer months was always planned
before 10:00 and after 18:00 so as to minimise sun glint. A polarised
lens Z neutral density (ND16) filter was also used. Wind speed and gust
measurements were recorded before and after each flight, through the
use of an AMES RVM 96B-1 handheld anemometer from Belfort
Instruments, as supplementary field data (mainly to test the drone
under different wind conditions and to respect maximum wind speed
thresholds [5 m/s] set by the insurance company).
Following data collection, the image sets were processed by the
OpenDroneMap software for the generation of a point cloud and texture
map (Open Drone Map, 2017). GPS information extracted from the
EXIF information of each image file was used for the creation of a
georeferenced orthophoto map having a resolution of 1 cm/pixel. This
was imported in Google Earth© and tiled for visualisation. Placemarks
were then manually added to all recorded litter items. In particular,
plastics were marked through a blue marker. Rope, wood and rubber
items were marked through green, red and black markers, respectively.
White was used for all other non-natural items. Different marker icon
polygons were used to represent different litter size categories. KML
files were also generated and imported in ArcGIS for the computation of
density maps. This made possible the identification of accumulated
litter hotspots.
To ensure the correct categorisation of debris, blind tests were
carried out by comparing the digitised data with field measurements
taken after each drone flight. Some objects that were manually identi-
fied (such as plastic bottles) were compared to the markers on Google
Earth to make sure that these were correctly marked and classified.
3. Results
Datasets from each survey at Qawra Point and Baħar iċ-Ċagħaq
generated 115 and 358 images, respectively. Fig. 3 visualises the digi-
tised database of the detected marine litter.
The total number of litter items recorded at the two monitored sites
(Bahar Ic-Caghaq and Qwara) was considerably different (30 items vs
A. Deidun et al.
Fig. 1. Map of the sampled Qawra Point (top) stretch at 35.959°N latitude 14.427°E longitude and the Baħar iċ-Ċagħaq (bottom) stretch at 35.944°N latitude
14.454°E longitude.
578 items, respectively), potentially as a consequence of differences
between the degree of wave and wind exposure of the two sites, as well
as between their coastal current dynamics, their coastal topography and
the degree of human activities on site. The highest concentration of
plastic litter was recorded along the western flank of Baħar iċ-Ċagħaq.
Successive surveys suggest an accumulation of wood debris on land as
well as of rope fragments close to the water. Only one rubber item was
found in the area. Other litter items which were observed over the
course of this study included chairs, mattresses, bed frame, and even a
washing machine. A total of 294 items was recorded for this area, with
the most frequently-recorded category being that of plastic items be-
longing to the > 50 cm category.
Along the east bay at Baħar iċ-Ċagħaq, while all plastic size cate-
gories were recorded, wood was found to be the predominant type of
debris. High concentrations were found inland. Small quantities of rope
fragments used for fishing were also recorded. The number of litter
items belonging to the ‘others’ category was found to increase with
time. A total of 284 litter items was recorded. Beached litter density
maps for this location are presented in Fig. 4.
Most of the debris items in Qawra Point were plastic fragments
shorter than 25 cm. Wood, rope, and other items were rarely observed.
A total of 30 items was recorded. Due to the low abundance of beached
Fig. 2. Drone transects implemented at Qawra Point (left) and at Baħar iċ-Ċagħaq (right).
214
Marine Pollution Bulletin 131 (2018) 212–217
litter items in this area, density maps were not produced. Fig. 5 sum-
marises the composition of beached litter recorded at the two coastal
study sites.
The computed statistics suggest that the results are not affected by
the day in which the survey is carried out. However, the concentration
and size parameters were found to depend on the location. Qawra Point
is a Blue Flag beach. Although data was collected as quickly as possible
after stormy days, the frequent manual cleaning of the area by workers
might have introduced a bias in the observations. Baħar iċ-Ċagħaq is
not protected nor manually cleaned. This might be the reason why
higher numbers are observed.
4. Discussion
To date, a number of different snapshot marine and coastal litter
assessments have been conducted in the Maltese Islands. Sciberras and
Axiak (1995) surveyed two beaches in Malta, where they recorded a
high presence of plastic and wood items. In a separate assessment of
litter along Maltese beaches, Tudor et al. (2002) concluded that this
mainly consisted of ‘beach user’ debris. This conclusion was further
supported by the Turner and Holmes (2011) study which confirmed
that litter on Maltese beaches mainly originated from land-based
A. Deidun et al. Marine Pollution Bulletin 131 (2018) 212–217

Fig. 3. Total marine litter detected on the west (left) and east (middle) bays in Baħar iċ-Ċagħaq as well as in Qawra Point (right). Key: blue = plastics; green = rope;
red = wood; black = rubber; white = all other non-natural items (e.g. fabric/textile). (For interpretation of the references to colour in this figure legend, the reader is
referred to the web version of this article.)

sources. Most recently (winter 2016 and summer 2017), within a
comprehensive survey along the entire 271 km-long coastline of the
Maltese Islands, Curmi and Axiak (2017) conducted observations on
beached litter by standing 1.5 m above the sea surface on a boat tra-
velling at constant speed along shore-parallel transects, in line with the
monitoring protocol for coastal litter described by Galgani et al. (2013)
for the MSFD. This study reports the accumulation of marine litter in
areas which are poorly urbanized and also during the winter season,
while, surprisingly, lower degrees of litter accumulation were recorded
in popular areas and during summer seasons, indicative of the mobility
of marine litter and the intrinsic influence of surface currents in moving
around such litter.
The total number of litter items recorded at the two monitored sites
– Qawra Point and Baħar iċ-Ċagħaq – was considerably different (30
items vs 578 items, respectively). Although such a stark difference
might be real and may be due to differences in the dynamics of coastal
currents and in coastal topography (e.g. profile), human activities
probably also play a role. Baħar iċ-Ċagħaq is in fact much more ac-
cessible from land to vehicles, making dumping of litter from the land
side a feasible prospect.
Indeed, at Baħar iċ-Ċagħaq, categories such as wood and plastic
were found on land, at considerable distances from the shoreline, near
points of vehicular access. It has been observed, in fact, that many
barbeques are held in this area and this could explain the accumulation
of wood items in the area, along with that of plastic bags containing
discarded food, further confirming the popularity of the site for re-
creational purposes.
The relative accumulation of rope fragments along the western side
of Baħar iċ-Ċagħaq could be indicative of the area's importance for
berthing purposes or its exposure to other important berthing and even

Fig. 4. Coastal litter density obtained for the west (left) and east (right) bays in Baħar iċ-Ċagħaq.

215
A. Deidun et al.
Fig. 5. variation in the abundance of litter items recorded along the western and eastern flanks of Bahar Ic-Caghaq B and at Qawra Point, during the summer 2017.
aquaculture (these sites deploy ropes on a large scale) areas further up
the coast, while the eastern flank of Baħar iċ-Ċagħaq is largely shielded
from north-western winds. On the other hand, considerably less litter
was recorded at Qawra Point, and this was mainly composed of plastic.
Since this place is near a touristic location (it is close to the Malta
National Aquarium), one can infer that probably the area is cleaned/
groomed more regularly than Baħar iċ-Ċagħaq, also in view of the Blue
216
Marine Pollution Bulletin 131 (2018) 212–217
Flag status of the Fra Ben sandy beach located close to Qawra Point. A
progressive accumulation of litter, especially in terms of rope and wood
fragments, within the monitored coastal sites was observed during
successive visits, suggesting a greater presence of such litter in the
marine domain with the progression of the summer season.
The innovative monitoring protocol presented in this study has both
a pronounced management and a policy relevance. In particular, the
A. Deidun et al.
data collection protocol represents a ready-to-use, off-the-shelf metho-
dology on how to deploy aerial drones for beached litter monitoring
given that several trials have been conducted within the current study
to optimise the same protocol. It has been observed in the current study
that a relatively large sampling area can be covered through drone
flights within a short period of time. The same flight path can be re-
produced faithfully on successive occasions due to geo-referencing,
making the monitoring programme a spatially-consistent one, taking
pictures every few seconds in order to record the entire perimeter of the
surveyed area.
Moreover, using Google Earth to geo-reference spotted Coastal
Litter further saves on photo analysis time and allows further manip-
ulation of the plotted items for further statistical analyses. This study
covered three coastal areas and it has been concluded in a couple of
months; therefore, if a more comprehensive drone-mediated monitoring
protocol were to be implemented all over the Maltese Islands, it would
necessitate the deployment of a number of drones which would cut
down greatly on operating costs of the monitoring (no boats and less
man hours needed). In this way, the protocol could prove useful to
coastal managers by identifying coastal litter accumulation areas, even
along stretches which are not easily accessible from land or by boat,
thus prioritising clean-up efforts. In particular, this approach appears
well suited to the monitoring needs of Marine Protected Area (MPA)
managers, enabling the regular and exhaustive monitoring of long
stretches of coastline within the confines of the same MPAs.
From a policy perspective, the innovative outputs of the current
study come at a very propitious timing given the imminent completion
of the first six-year cycle of the MSFD (Marine Strategy Framework
Directive), which was scheduled to span over the 2012–2018 period.
The current stage within the MSFD implementation timeframe presents,
in fact, a window of opportunity for innovative monitoring solutions
which might be adopted within revised monitoring protocols for the
second MSFD six-year cycle if their scientific rigour and applicability is
proven. Quantitative (including individual litter item dimensions) and
qualitative (including identification and categorisation of individual
litter items) data gleaned from the aerial imagery collected within this
study compared well with corresponding ground truthing data collected
by workers in the field, thus making the mainstreaming of drone-
mediated monitoring within national and regional monitoring pro-
grammes for marine litter more plausible.
According to a recent European Commission's assessment report to
the EU Parliament and Council on the MSFD monitoring programmes
formulated by Member States (EU Commission, 2017), although almost
all MS have developed programmes to monitor marine litter and there
is appropriate spatial coverage and frequency of monitoring for beach
litter, there are still gaps, mainly related to litter monitoring at the
water surface. Aerial drones operated from vessels, especially those
with an enhanced battery autonomy, can assist in bridging this gap, and
these can be supplemented by the information from underwater drones
deployed from vessels as they become more and more commercially
available.
In a review issued by the Joint Research Centre (JRC) of the EU
Commission's decision concerning MSFD criteria for assessing Good
Environmental Status (GES), the replacement of the term ‘possible’ with
‘feasible’ in proposed within the text for Indicator 10.1.1 (‘Trends in the
amount of litter washed ashore and/or deposited on coastlines, in-
cluding analysis of its composition, spatial distribution and, where
feasible, source’). As amply demonstrated in the current study, aerial
imagery can feasibly lead workers to infer the source of beached litter
by expanding spatially the monitoring perimeter, both inland and
seaward, and through an integration with hydrodynamic and meteor-
ological information. Comprehensive remote data portals, such as the
online Copernicus one (www.copernicus.eu) for instance, provide daily
values for water currents at the surface, mid-column and above the
seabed, besides mean wind speed and direction, with hindcast and
forecast capacities provided through mathematical modelling.
217
Marine Pollution Bulletin 131 (2018) 212–217
Despite the potential advancement of beached litter monitoring
capacities through the deployment of aerial drones, possible constraints
imposed by data protection considerations, which have resulted in
many countries adopting restrictions on drone flights in some areas
(including beaches), need to be factored in. Inclement weather condi-
tions, especially strong wind speeds, are yet another constraint for this
novel monitoring protocol. It is difficult to discriminate, solely through
an examination of drone footage, between the plastic litter deposited
along the coast by the sea and litter recorded further inland within the
coastal region and having a terrestrial region (i.e. through human
visitors to the area). This can tentatively be achieved through a direct
examination in the field of litter to assess parameters such as degree of
degradation, since sea-borne litter is presumably subjected to higher
rates of weathering.
Acknowledgements
The authors are grateful to the Research Committee of the
University of Malta for supporting financially the study reported herein
through a grant allocated in February 2017.
Conflicts of interest
None.
References
Capolupo, A., Pindozzi, S., Okello, C., Boccia, L., 2014. Indirect field technology for de-
tecting areas object of illegal spills harmful to human health: application of drones,
photogrammetry and hydrological models. Geospat. Health 8 (3), 699–707.
Casella, E., Rovere, A., Pedroncini, A., Stark, C.P., Casella, M., Ferrari, M., Firpo, M.,
2016. Drones as tools for monitoring beach topography changes in the Ligurian Sea
(NW Mediterranean). Geo-Mar. Lett. 36 (2), 151–163.
Cózar, A., Echevarría, F., González-Gordillo, J.I., Irigoien, X., Úbeda, B., Hernández-León,
S., Palma, A.T., Navarro, S., Garcia-de-Lomas, J., Ruiz, A., Fernández-de-Puelles,
M.L., Duarte, C.M., 2014. Plastic debris in the open ocean. Proc. Natl. Acad. Sci. 111
(28), 10239–10244.
Curmi, M., Axiak, V., 2017. Levels of floating litter in Maltese coastal waters (Central
Mediterranean). In: Ozhan, Malta E. (Ed.), Proceedings of the 13th MEDCOAST
International Congress on Coastal and Marine Sciences, Engineering, Management
and Conservation MEDCOAST 2017, 31st October–4th November 2017, Mellieha.
Eriksen, M., Lebreton, L.C., Carson, H.S., Thiel, M., Moore, C.J., Borerro, J.C., Galgani, F.,
Ryan, P.J., Reisser, J., 2014. Plastic pollution in the world's oceans: more than 5
trillion plastic pieces weighing over 250,000 tons afloat at sea. PLoS One 9 (12),
e111913.
European Commission, 2017. Report From the Commission to the European Parliament
and the Council Assessing Member States' Monitoring Programmes Under the Marine
Strategy Framework Directive. Brussels. 16.01.17. (15pp).
Galgani, F., Hanke, G., Werner, S.D.V.L., De Vrees, L., 2013. Coastal litter within the
European marine strategy framework directive. ICES J. Mar. Sci. 70 (6), 1055–1064.
Ioakeimidis, C., Zeri, C., Kaberi, H., Galatchi, M., Antoniadis, K., Streftaris, N., ...
Papatheodorou, G., 2014. A comparative study of marine litter on the seafloor of
coastal areas in the Eastern Mediterranean and Black Seas. Mar. Pollut. Bull. 89 (1–2),
296–304.
Lebreton, L.M., Greer, S.D., Borrero, J.C., 2012. Numerical modelling of floating debris in
the world's oceans. Mar. Pollut. Bull. 64 (3), 653–661.
Linchant, J., Lisein, J., Semeki, J., Lejeune, P., Vermeulen, C., 2015. Are unmanned air-
craft systems (UASs) the future of wildlife monitoring? A review of accomplishments
and challenges. Mammal Rev. 45 (4), 239–252.
Open Drone Map, 2017. Computer Software. Retrieved from. https://github.com/
OpenDroneMap/OpenDroneMap.
Sciberras, C., Axiak, V., 1995. Coastal marine litter in the central Mediterranean: Baseline
information on beach stranding, coastal densities and rates of photodegradation.
CIESM. Rapp. Comm. int. Mer Medit. 34, 146.
Suaria, G., Aliani, S., 2014. Floating debris in the Mediterranean Sea. Mar. Pollut. Bull. 86
(1), 494–504.
Tudor, D.T., Williams, A.T., Randerson, P., Ergin, A., Earll, R.E., 2002. The use of mul-
tivariate statistical techniques to establish beach debris pollution sources. J. Coast.
Res. 36, 716–725.
Turner, A., Holmes, L., 2011. Occurrence, distribution and characteristics of beached
plastic production pellets on the island of Malta (central Mediterranean). Mar. Pollut.
Bull. 62 (2), 377–381.
UNEP/MAP, 2015. Marine Litter Assessment in the Mediterranean. Athens, Greece.
(145pp. ISBN No: 978 -92 -807 -3564 2).
Zhang, J., Hu, J., Lian, J., Fan, Z., Ouyang, X., Ye, W., 2016. Seeing the forest from
drones: testing the potential of lightweight drones as a tool for long-term forest
monitoring. Biol. Conserv. 198, 60–69.