Embed Size (px)
Citation preview
A Global Platform for Monitoring Marine Litter and Informing Action
Draft version: 5 December 2019
Emily Smail, Daniel Takaki, Hans-Peter Plag, René Garello, Samy Djavidnia, José Moutinho, Ghada El Serafy, Alessandra Giorgetti, Matteo Vinci, Maria Eugenia Molina Jack, Kate Larkin, Dawn J. Wright, Francois Galgani, Konstantinos Topouzelis, Stewart Bernard, Anne Bowser,
Guido Colangeli, Linwood Pendleton
Table of Contents
Table of Contents .......................................................................................................................... 2
Background and context ............................................................................................................... 1
Section 1: Monitoring technologies ............................................................................................... 4
Section 2: Existing marine litter databases and major datasets ................................................. 19
Section 3: Indicators and Applications of Technologies .............................................................. 29
Section 4: Existing and Developing Global Data Platforms ........................................................ 40
Section 5: Proposed Features of a Global Platform and Required Resources ........................... 48
Section 6: Marine Litter in a Digital Ecosystem for the Environment – Thought on A Pilot Project .................................................................................................................................................... 54
Section 7: Future Developments ................................................................................................. 57
References .................................................................................................................................. 59
1
Background and context 1
Marine litter is defined as “any persistent, manufactured or processed solid material discarded, 2 disposed of or abandoned in the marine and coastal environment”. Marine litter is found in all the 3 world's oceans and seas, even in remote areas far from human contact due to its transboundary 4 nature. The continuous growth in the amount of solid waste thrown away and the slow rate of 5 degradation of most items are together leading to a gradual increase in marine litter found at sea, 6 on the sea floor and coastal shores. It has become an economic, environmental, human health 7 and aesthetic problem posing a complex and multi-dimensional challenge. 8
Marine plastics are of particular interest due to the fact that in the last 50 years, plastic production 9 has increased more than 22-fold while the global recycling rate of plastics in 2015 was only an 10 estimated 9% (Geyer et al., 2017). This rise in plastic production and unmanaged plastic waste 11 has resulted a growing threat to marine environments with an estimated 5-13 million tons of plastic 12 from land-based sources ending up in marine environments (Jambeck et al., 2015). 13
The importance of marine plastics has been recognized in the Sustainable Development Goals 14 through a target related to marine litter (SDG target 14.1) and in four UN Environment resolutions 15 (from UNEA-1 in 2014, UNEA-2 in 2016, UNEA-3 in 2017 and UNEA-4 in 2019). However, there 16 are large gaps in knowledge in terms of understanding marine litter and microplastics: a reliable 17 figure for the volume of plastics entering the ocean, the accumulated volume of plastics in the 18 marine environment, mapping of the source and sink location of plastics and basic data on 19 microplastic is currently lacking. There is a need to use existing data from remote sensing, citizen 20 science and in situ monitoring to better understand marine litter and microplastics; however, much 21 of the research in this field is at an initial stage and only data related to beach litter is available in 22 many regions(UN Environment, 2018). 23
A consistent need for monitoring and reporting of marine litter has been recognized in the 24 Sustainable Development with Goal 14, Target 14.1: “by 2025, prevent and significantly reduce 25 marine pollution of all kinds (…)”. This target provides a deadline for progress on reducing marine 26 litter which is informed by SDG indicator 14.1.1b, “marine litter.” UN Environment is proposing 27 four sub-indicators for SDG 14.1.1b: 28
1) Plastic debris washed/deposited on beaches or shorelines (beach litter) 29
2) Floating plastic debris and debris in the water column 30
3) Plastic debris on the seafloor/seabed 31
4) Plastic ingested by biota (e.g. sea birds). 32
Despite the growing interest in monitoring the above areas, there is a wide range of non-33 comparable monitoring approaches that limits the development of indicators and spatial or 34 temporal assessments (Galgani, Hanke & Maes, 2015). Additionally, these measures only 35 capture the accumulation of plastics and do not cover marine litter more broadly, do not cover 36 microplastics and do not cover the sources and pathways for marine litter. In order to effectively 37 monitor marine litter there is a need to also consider the following: 38
2
- Plastic flow: How marine litter moves in the marine environment in a way that allows 39 tracking the origin of plastic pollution is 40
- Life-cycle approach/waste/recycling/reduction: Monitoring should encompass not only 41 monitoring of plastic in the ocean, but across the life-cycle. This requires a holistic 42 approach which would entail monitoring production practices across the life cycle of 43 products as well as how plastic becomes waste. 44
- Waste management practices: Leakages in the waste management system, illegal 45 dumping and leakages in recycling are a leading contributor to marine litter. Waste 46 management is also captured in the SDGs and is an import contributor to marine litter. 47
- Plastic in waterways: There is a lack of information on how plastic and microplastic 48 move through rivers, sewage systems and other waterways to end up in the marine 49 environment. 50
- Plastic types: Tracking of plastics and microplastics by type of plastic, including plastic 51 related e-waste and chemicals (and toxicity) in plastics, is needed. Without information on 52 what can be recycled and what has chemicals then it is not possible to get a complete 53 picture. 54
- Consumer awareness: Information and data must be communicated in a way that builds 55 awareness of the public so that consumers can make informed decisions. 56
- Microplastics: Understanding the sources of microplastics and the impact of 57 microplastics on human health is a priority. 58
- Trade-offs: Understanding when the alternative to plastic is worse than the use of plastics 59 is a challenge. This makes providing policy advice that would result in benefits and not 60 costs to human health and the environment difficult. 61
Marine litter observation is currently very sparse, and as a result, very little is known in particular 62 about the biological and physical process that transport plastics through marine ecosystems and 63 potentially to humans (Katija et al. 2017). Therefore, any quantitative approach to integrating 64 source and dispersion/accumulation dynamics must take a multidisciplinary approach combining 65 forward or inverse hydrodynamic or dispersion models with multisource Earth observation 66 observation data. 67
While standardized methods for monitoring marine litter will greatly improve the understanding of 68 the marine litter, development and reporting of indicators will require integrated and comparable 69 data. Currently, much of the data on marine litter is published in peer-reviewed journals and 70 distributed in databases hosted by NGOs and government authorities. As suggested by Galgani, 71 Hanke & Maes, 2015, and Maximenko et. al 2019, a joint international database would facilitate 72 the collection of data for marine litter pollution indicators and improve standardization and 73 comparability. Such a database would also support policy decisions related to the reduction of 74 marine litter and support analysis of the efficacy of mitigation efforts. 75
This paper outlines a concept for the development of a global data platform for marine litter 76 including vision, feasibility and potential structure and funding needed. This paper will be used to 77 further discussions with the view to develop a long-term project in support of such a platform 78 which could be hosted on the UNEP World Environment Situation Room. This paper is organized 79 into five sections: 80
Section 1 provides a summary of existing and developing monitoring technology. 81
3
Section 2 provides a summary of existing marine litter databases and major published datasets. 82
Section 3 outlines existing and proposed indicators based on available technologies and data. 83
Section 4 provides a summary of existing and developing platforms of relevance. 84
Section 5 outlines the proposed features of a global platform for monitoring marine litter and 85 informing action, next steps and required resources. 86
Section 6 outlines a proposed pilot project for the development of marine litter in a digital 87 ecosystem for the environment. 88
Section 7 provides insights into aspirational, future developments. 89
90
91
4
Section 1: Monitoring technologies 92
Regular and standardized monitoring of marine litter is needed in order to understand long-term 93 changes in marine litter and for the successful development and implementation of mitigation 94 strategies. The diverse nature, sources and impacts of marine litter require a wide range of 95 technologies and methods for monitoring. Recent efforts to compile information on existing 96 methodologies and recommend standardized methodologies for global monitoring include the 97 Joint Group of Experts on the Environmental Aspects of Marine Environmental Protection 98 (GESAMP) Guidelines for the monitoring and assessment of plastic litter and microplastics in the 99 ocean (GESAMP, 2019) and the Global Manual on Ocean Statistics (UN Environment, 2018). 100
One challenge for implementing and further developing monitoring methodologies for marine litter 101 is an understanding of the existing technologies for monitoring marine litter. This section 102 summarizes technologies for the monitoring of marine litter and how these technologies can be 103 used for collecting the necessary data for a global view of marine litter. For an overview of the 104 observing system technologies required for the development of a future integrated marine debris 105 observing system, see (Maximenko et. al. 2019). 106
For this paper, we have grouped technologies based on applicability to the size classes 107 recommended in the GESAMP 2019 methodology (Table 1). In addition, we have assigned 108 technology readiness levels (Table 2) based on the National Oceanic and Atmospheric 109 Administration (NOAA) policy on research and development transitions to support prioritization of 110 data standardization and integration. Readiness levels are defined by NOAA as “A systematic 111 project metric/measurement system that supports assessments of the maturity of research and 112 development projects from research to operation, application, commercial product or service, or 113 other use and allows the consistent comparison of maturity between different types of research 114 and development projects” (NOAA, 2017). 115
Table 1. Size categories for routine marine litter monitoring (GESAMP, 2019)
Size Category Size Range Mega > 1 m Macro 25 mm – 1 m Meso 5-25 mm Micro <5 mm
Table 2. Technology readiness levels (NOAA, 2017)
Readiness Level Readiness Level Defined 1 Basic research and/or development principles observed and
reported 2 Formulation of concept for operations, application,
commercialization or other uses for societal benefits 3 Proof-of-concept (viability established) 4 Validation of system, process, product, service or tool in
laboratory or other experimental environment 5 Validation of system, process, product, service or tool in
5
relevant environment 6 Validation of system, process, service, or tool in relevant
environment (potential demonstrated) 7 Prototype demonstrated in an operational or other relevant
environment (functionally demonstrated in pseudo real world environment)
8 System, process, product, service, or tool completed and “mission qualified” through test and demonstration in operational or other relevant end-to-end environment (functionality demonstrated)
9 System, process, product, service or tool approved for deployment and use in decision making (transition complete)
Human observers 116
Visual human observation the most wide spread and technically simplistic way to collect data 117 about marine litter and is used for the monitoring of beach/shoreline litter, floating litter, water 118 column litter, seabed/seafloor litter, marine litter ingestion/entanglement and sources of marine 119 litter. Human observation is most appropriate for macro- and mega-litter based on what is 120 consistently visible to the naked eye (GESAMP, 2019). 121
Protocols and guidelines for monitoring beach/shoreline litter with visual observations varying 122 widely by organization (Arctic Council, 2015; Cheshire et al., 2009; European Commission JRC, 123 2013; NOWPAP CEARAC, 2007; Opfer et al., 2012; UN Environment, 2016). For litter on the 124 beach/shoreline surface, analysis is typically done through done through visual transects and 125 accounting of collected items from beach cleanup efforts. Some organizations employ apps to 126 facilitate data entry and reporting (e.g., NOAA Marine Debris Tracker App1, European 127 Environment Agency’s Marine LitterWatch App2, Ocean Conservancy’s Clean Swell App3). 128
Floating litter monitoring by human observers is typically done using transects from ships. While 129 differences in methods are used, visual surveys from ships for floating marine litter have been 130 used for almost 50 years and is an important source of data (GESAMP, 2019). These 131 observations are generally limited to mega- and macro-litter. Human observations of water 132 column litter require collection of materials with net tows with collected samples and can be used 133 for analysis of meso-litter. Items collected from net tows are typically sorted by type and size and 134 analyzed by count and/or weight (GESAMP, 2019; L. Lebreton et al., 2018). 135
In shallow waters, monitoring and collection efforts can be carried out via underwater visual 136 surveys by SCUBA divers. Marine litter density is measured most commonly by distance and 137
1https://marinedebris.noaa.gov/partnerships/marine-debris-tracker
2https://play.google.com/store/apps/details?id=com.litterwatch&hl=en_US
3https://www.coastalcleanupdata.org/#download
6
transect sampling (Buckland, 2001; Galgani et al., 2013; Spengler, 2008). This method is limited 138 in its depth typically to 20-30 m at most and required SCUBA equipment and skilled observers 139 and is most appropriate for macro-litter. In addition to professional surveys, recreational divers 140 also play a role in surveys. For example, divers through Project AWARE’s Dive Against Debris 141 program are encouraged to collect and report marine litter found underwater, and they are 142 directed to collect and observe at the same locations when they do for further data validation 143 (GESAMP 2019). 144
Visual reporting of abandoned, lost, or otherwise discarded fishing gear is an important part of 145 monitoring entanglement and entanglement risk. Overall, monitoring entanglement has a 146 straightforward observational methodology, where it is important to note the size, location, 147 impacted species or habitat, as well the type of litter when reporting on entanglement (GESAMP 148 2019). Networks for reporting of entanglement and litter with entanglement risk include the NOAA 149 SOS Whale Network4 and the International Association of Geophysical Contractors Marine Debris 150 and Ghost Net Initiative5. As an example, a dedicated study conducted between 2003 and 2015 151 on pollution incidents reported by the observers on-board fishing vessels in the Western and 152 Central Pacific Ocean (Richardson et al., 2017), showed that 71 - 80% percent of the incidents 153 reported were documented as waste dumped overboard, and only 13 -17 % as abandoned, lost 154 or dumped fishing gear, depending on the type of vessel. Increased observer coverage and data 155 collection on-board ships such as fishing vessels may provide essential information. Extending 156 such observations to other vessels would provide more information about the quantities and types 157 of pollution caused by shipping. Navigation logs are designed for reporting of pollution incidents 158 on-board and this would continue to be an appropriate form for use by an expanded, cross-fleet 159 observer program, further organized, quality controlled and standardized to Global information 160 Systems (IMO, GOOS, etc.). 161
Human observations are used for monitoring sources of marine litter including floating riverine 162 inputs and leakage from waste sites. In Europe, the Riverine Litter Observation Network6 uses 163 human observation of floating macro litter on the river surface. An added challenge to visual 164 observations of floating litter in riverine environments include surface water speed and turbulence 165 (González-Fernández & Hanke, 2017). In order to establish and estimate the linkage between 166 land-based waste management and losses of waste into the marine environment, human 167 observers are used to conduct terrestrial litter surveys of inland, riverine and coastal areas 168 (Schuyler et al., 2018). 169
Data collected from human observers has been test and used extensively for analysis in regions 170 including the North-East Atlantic, Baltic Sea and United States (European Commission JRC, 171 2013; Hardesty et al., 2017; OSPAR, 2017). However, since standardized global protocol/process 172
4https://www.westcoast.fisheries.noaa.gov/protected_species/marine_mammals/disentanglement_network.html
5https://www.iagc.org/ghost-net-contact-form.html
6https://mcc.jrc.ec.europa.eu/main/dev.py?N=simple&O=394&titre_page=RIMMEL%2520observation%2520Network
7
for collecting data using human observations has not yet been implemented, we have assigned 173 a readiness level of 8. 174
Readiness Level: Human Observers
8: System, process, product, service or tool completed and “mission qualified” through test and demonstration in operational or other relevant end-to-end environment (functionality demonstrated)
Microscopy 175
Microscopy is used primarily for analysis of meso- and micro-litter and has applications for the 176 monitoring of beach/shoreline litter, floating litter, water column litter, seabed/seafloor litter, 177 marine litter ingestion and sources of marine litter. Sample collection for beach/shoreline litter is 178 typically done by collecting sediment with a spoon, spoon trowel or sediment core and passing 179 the sample through various sieves depending on the size class of interest (GESAMP, 2019). 180 Floating/water column samples require filtration either after the samples are collected or using in 181 situ filtration equipment (Choy et al., 2019; GESAMP, 2019). Samples for ingestion are typically 182 taken from dead organisms or from items associated with live animals such as regurgitated 183 pellets, scat and nesting materials (GESAMP, 2019). In addition, submersible microscopes e.g. 184 holographic (4deep) or cytometric, can be placed in typical outflow areas to autonomously 185 measure micro-plastics. The use of digital holographic microscopy, matched with the continuous 186 advancements in deep learning techniques can provide new opportunities for the use of coherent 187 imaging systems in many areas, including potentially microplastic pollution analysis (Rivenson et. 188 al., 2019) 189
Methods for sample preparation and analysis vary widely based on sample type (e.g. water 190 sample, sediment sample, ingested sample) and microscopy type (e.g. light microscopy, electron 191 microscopy, etc.). Microscopic analysis is often targeted solely on micro-plastics. To ensure all 192 organic matter is removed from samples prior to analysis of microplastics, several chemical 193 digestive methods have been utilized. These methods are broken down into three general 194 categories: oxidative, acidic, alkaline/basic and enzymatic (table 4). 195
Table 3. Advantages and disadvantages for extracting and purifying microplastics in organic matrices (GESAMP, 2019)
Purification Method Advantages Disadvantages Oxidative Digestion • Inexpensive • Temperature needs to be
controlled • Several applications may be
needed Acid Digestion • Rapid (24 hr) • Can attack some polymers Alkaline Digestion • Effective
• Minimal damage to most polymers
• Damages cellulose acetate
Enzymatic Digestion • Effective • Minimal damage to most
• Time-consuming (several days)
8
polymers
Methods for extraction of ingested litter from samples vary widely (Courtene-Jones et al., 2019; 196 GESAMP, 2019; van Franeker et al., 2011; Zhao et al., 2016) and need a standardized approach 197 to ensure consistency. 198
Another challenge for analysis of marine litter by microscopy is the potential for sample 199 contamination. Careful procedures to avoid sample contamination during analysis are being 200 implemented in research studies such as burning off contaminants from glassware, pre-filtering 201 of reagents through glass fiber filters, handling of samples in laminar flow hoods and analysis of 202 blanks to estimate potential contamination (GESAMP, 2019; Wesch et al., 2017; Zhao et al., 203 2016). 204
Analysis by light microscopy typically consists of counting microplastics and characterizing their 205 color, shape and sizes (Vandermeersch et al., 2015). Scanning electron microscopy can provide 206 additional detail about the surface texture of particles but is only viable for analysis of small 207 quantities of samples do to the intensive processing and analysis required (GESAMP, 2019). 208 Overall, various microscopic approaches have trade-offs in terms of precision and accuracy of 209 material identification with some methods potentially underestimating microplastic pollution due 210 to false positives (Zarfl, 2019). As protocols for cleanly and accurately collecting, processing and 211 analyzing samples for microscopy are still being researched, we have assigned a readiness level 212 of 3 to light microscopy. 213
Readiness Level: Microscopy
3: Proof-of-concept (viability established)
Weighing 214
Technology for calculating the mass of marine litter is frequently used for the analysis of macro, 215 meso and micro litter beach/shoreline litter, floating litter, water column litter, seabed/seafloor 216 litter, marine litter ingestion/entanglement and sources of marine litter (L. Lebreton et al., 2018) 217 (L. C. M. Lebreton et al., 2017; NOAA Marine Debris Program, 2015). Mega-debris is difficult to 218 weigh, compounded by the fact that mega-debris such as fishing nets are often heavily colonized. 219 Challenges for accurately weighing larger items include sand or debris entangled in the item and 220 consistency in properly drying samples (GESAMP, 2019). Technologies for weighing macro and 221 meso litter tend to be simple including scales and drying ovens (Ryan et al., 2014). 222
Accurate mass (or gravimetric analysis) of micro-plastic requires proper sorting, extraction and 223 sample purification as outlined in the light microscopy section. Consistency in mass 224 measurements have shown to be consistent across labs when the same method is applied for 225 analysis.(NOAA Marine Debris Program, 2015). One critical aspect to consider for mass 226 calculation is that most methods for microplastic analysis include an density separation step 227 where settled solids are discarded and only floating solids are analyzed (GESAMP, 2019; NOAA 228 Marine Debris Program, 2015). One consideration regarding this approach is that scanning 229 electron microscopy images have shown microplastics to have extensive fouling by microbial 230 communities (Zettler et al., 2013) which can cause plastic debris to sink (Andrady, 2011). 231
9
Accordingly, proper steps need to be taken prior to analysis of the weight of micro-litter to ensure 232 that biofouling does not result in an underestimate. 233
As standardized approaches to removing sand, biofouling and water residue from samples have 234 not been implemented and methods often vary widely or are not specifically reported, we have 235 assigned a technology readiness level of 3 to weighing marine litter. 236
Readiness Level: Weighing Litter
3: Proof-of-concept (viability established)
Spectroscopy 237
The analysis of absorption or scattering of light via spectroscopy allows for the discrimination 238 between organic and inorganic particles as well as various types of plastics as these materials 239 produce different spectral signals (Lenz et al., 2015). While the waste management and recycling 240 industries have utilized near-infrared spectroscopy to identify plastics since 1998, the utilization 241 of spectroscopy for analysis of marine litter is fairly recent (Choy et al., 2019; Yu et al., 2019; Zhu 242 et al., 2019; Zulkifley et al., 2014). 243
Spectroscopy techniques such as Fourier Transform Mass Spectroscopy (FTIR) and Laser Raman 244 Spectroscopy have been primarily focused on analysis of microplastics in the marine environment 245 (Choy et al., 2019; GESAMP, 2019; Yu et al., 2019). New spectroscopic approaches such as staining 246 and semi- or fully automated spectroscopic analysis are currently under development (GESAMP, 247 2019). As Sample collection, treatment and analysis methods are still in the research and development 248 phase for the identification of marine plastics by spectroscopy, we have assigned a readiness level of 249 1 to spectroscopy. 250
Readiness Level: Spectroscopy
1: Basic research and/or development principles observed and reported
Mass Spectronomy 251
Mass spectronomy measures the mass to charge ratio of ions in a sample, providing information 252 about chemical composition. Mass spectronomy technologies that have been utilized for analysis 253 of microplastic particles include thermal extraction and desorption gas chromatography mass 254 spectrometry (TED-GC-MS) and pyrolysis gas chromatography mass spectrometry (Py-GCMS). 255 These technologies require the thermal degradation of plastics, separation of degradation 256 products through chromatography and analysis of the products with mass spectronomy 257 (Dumichen et al., 2017; GESAMP, 2019). Other forms of (GESAMP, 2019) mass spectronomy 258 have been used to identify chemicals associated with plastic samples (Kuhn et al., 2018). For 259 example, Inductively coupled plasma spectroscopy (ICP-MS) can identify metals associated with 260 plastics which can provide information about hazardous metals associated with microplastics 261 (Kuhn et al., 2018). As the application of mass spectronomy to the analysis of marine litter is still 262
10
in the research and development phase, we have assigned a readiness level of 1 to mass 263 spectronomy. 264
Readiness Level: Mass Spectronomy
1: Basic research and/or development principles observed and reported
Visual Imagery and Video 265
Visual imagery is collected done using ship based cameras, unmanned aerial vehicles (UAVs), 266 balloons, high altitude pseudo satellites (HAPS), remotely piloted aircraft systems (RPAS) and 267 satellites. Both, drones and blimps are limited in terms of timing and spatial coverage, and their 268 use should be guided by model studies. Fixed-wing drones are increasing the distance and 269 duration of drone flights. Blimps have the advantage of longer stable flights. 270
The use of small aircraft, drones, unmanned aerial vehicles (UAVs), balloons, drones and 271 satellites are promising for the analysis of beach litter as well as sea surface litter. The advantages 272 of aerial technologies include access to imagery from difficult to access beaches, more rapid and 273 complete beach coverage and high-resolution imagery. Aerial imagery can be processed 274 manually or automatically using machine learning tools that are currently in development (Deidun 275 et al., 2018; Martin et al., 2018; Moy et al., 2018). The various aerial technologies have strengths 276 and weaknesses based on cost and coverage. For example, UAVs offer ultra-high resolution 277 imagery but are prohibited to fly over people, limiting survey locations (Moy et al., 2018). 278 Validation of results using ground measurements is an important component for the development 279 of these technologies and further tests will be needed to understand the limitations and 280 appropriate applications of aerial technologies for monitoring of beach litter (Deidun et al., 2018; 281 Moy et al., 2018). 282
For marine floating marine litter, photographing marine litter using a camera fixed to the bow or 283 mast of a vessel is an emerging approach for monitoring marine litter [additional information]. High 284 resolution cameras or other sensors (e.g. Lidar) mounded on ships can increase the observations 285 on the floating litters and with the use of AI, provide in situ observation in real time. Consistency 286 of such sensors is of great importance and has to be validated with further testing. In addition to 287 ship-based cameras, autonomously operated vehicles (AOVs) are being explored for the use of 288 surface/subsurface marine litter at sea. For example, Wave Gliders that use wave energy for 289 propulsion often have video cameras that can be used for marine litter quantification (Galgani et 290 al., 2013). 291
One way marine litter is monitored is via remotely operated vehicles such as submarines or 292 manned submarines which can view seabed litter plastic, or take core or surface samples to detect 293 presence of microplastics and other litter (Woodall et al., 2014). ROVs are often preferable when 294 litter surveys are conducted on continental slopes, uneven terrain, or the deep seafloor. Litter can 295 accumulate in certain locations on the seafloor such as coastal canyons, as well as areas with 296 steep slopes, rocky bottoms, or ocean trenches. These are areas that would specifically benefit 297 and often necessitate the use of ROVs to observe and/or collect marine litter. High resolution 298 images can be recorded with video cameras while other light devices such as lasers are used to 299 measure both transect areas as well as object size and distances on the seafloor. Proposed 300 learning algorithms aimed at more successful vision detection of litter will be useful in helping to 301
11
exploring and mapping litter by the way of autonomous underwater vehicles. While ROVs have 302 proven useful, the high cost of operation as well as the specific skill set required for both operation 303 and observation, remain a limitation. 304
There are various technologies to estimate riverine sources directly with varying levels of effort, 305 scale and accuracy. Drone or field surveys of river mouths can assess accumulated plastic. A 306 standard methodology for drone-based surveys of plastic is being developed by DRONET7. 307
Visual imagery and video have been utilized in relevant environments but do need 308 standardization. Accordingly, we have assigned a readiness level of 5 to the analysis of marine 309 litter by visual imagery and video. 310
Readiness Level: Visual Imagery and Video
6: Validation of system, process, service or tool in relevant environment (potential demonstrated)
Synthetic Aperture Radar 311
Synthetic Aperture Radar (SAR) has potential to provide information about expected locations of 312 marine litter and for the detection of mega-litter on the water’s surface and on beaches. Synthetic 313 Aperture Radar (SAR) provides high-resolution information on parameters of the ocean surface 314 such as topography, roughness, surface waves, winds and currents, which can be correlated with 315 the movement of marine debris, or used for the identification of convergent fronts where floating 316 litter collects. SAR can be used to locate proxies where litter is expected to be present i.e. 317 conversion zones and windrows in large areas. Unmanned aerial vehicle (UAVs) and high 318 resolution SAR satellite can be additionally used to verify the presence of litter. SAR sensors are 319 mounted on airborne platforms, such as the NASA L-Band Uninhabited Aerial Vehicle Synthetic 320 Aperture, and satellites. Larger objects, which float at the surface can be detected directly with 321 SAR or through secondary surface wave patterns, although false alarms remain a significant 322 obstacle (Davaasuren et al., 2018). As the use of SAR for marine litter monitoring is still in the 323 research and development phase, we have assigned a readiness level of 1 to SAR. 324
Readiness Level: Synthetic Aperture Radar
1: Basic research and/or development principles observed and reported
Multispectral and Hyperspectral imaging 325
Satellite remote sensing of beach litter and sea surface litter is currently in the research and 326 development phase, primarily repurposing missions that were not originally designed for litter 327
7https://www.theplastictide.com/blog-1/2018/4/22/launching-the-marine-litter-dronet
12
monitoring. Satellite imagery relevant for remote sensing of beach litter includes visual imagery 328 and spectral analysis. Commercial satellite imagery is the primary technology of relevance when 329 detecting litter on beaches and rivers, given the very high-resolution needed to discern this litter. 330 For spectral analysis, research activities to map the spectral signatures of marine plastics are 331 underway and show promise for potential characterization of marine litter on beaches (Acuna-332 Ruz et al., 2018; Garaba & Dierssen, 2018). Preliminary studies have shown the synergetic use 333 of satellite images and UAVs to detect floating litter (Topouzelis et al, 2019). Current high-334 resolution satellite sensors can be used for monitoring floating mega litters. Statistical indicators 335 and density heat maps can be derived in accordance to predefined requirements. Future satellite 336 sensors may show improved functionality for measuring marine litter on beaches but these 337 concepts are still in development. Given the current state of technology and applications research, 338 we have assigned a technology readiness level of 1 for satellite remote sensing of beach and sea 339 surface litter. 340
Multi-spectral satellite remote sensing of plastic in the water column is currently only possible for 341 larger elements on or close to the water surface, and under good atmospheric conditions (no 342 clouds). The Copernicus Sentinel 2 constellation is likely to be the most valuable existing mission, 343 with freely available data, relatively high resolution (10m GSD) spectral radiometry with global 344 coverage. Commercial higher (i.e. very high-resolution) satellite data are available for purchase 345 but have low temporal resolution. In all cases cloud cover and sea surface conditions affect the 346 detection of debris no matter the resolution. An initial assessment of observation requirements for 347 measuring marine plastic debris from space has been documented and can be used for further 348 sensors development (Martínez-Vicente et al, 2019). 349
As the plastic elements sink or decompose, the likelihood of detection with remote sensing 350 methods decreases significantly. There are some promising methods looking at anomalies or 351 particular signatures to identify ocean plastic. For example ESA’s Sentinel-3 satellite has an 352 ocean color imager that is potentially detecting unique signatures or large agglomerations of 353 plastic. However, this sensor only images at 300 m resolution, and even with a revisit rate of 354 almost every 2 days, it will not detect most plastic of interest. Commercially available 355 hyperspectral sensors such as HyMap may be more suited for detecting plastics (Garaba et al., 356 2018; Goddijn-Murphy et al., 2018). 357
As multispectral and hyperspectral imaging is still in the research and development phase, we 358 have assigned a readiness level of 1. 359
Readiness Level: Multispectral and Hyperspectral imaging
1: Basic research and/or development principles observed and reported
GPS tags and transmitters 360
Direct tracking of floating marine items can be achieved more reliably with debris tagged with 361 GPS tags and transmitters. Trajectories of marine litter can then be compiled to reconstruct the 362 path of plastic from source to fate. Tracking can be achieved with Argo tracking sensors, or GPS 363 devices, which may however remain too expensive to implement widely. The upcoming Kineis 364
13
constellation from CLS89, or a low-tech solution such as PandaSat10 proposed by WWF could 365 provide more affordable solutions in 2021. Using satellite trackers deployed on vessels of 366 opportunity in the Pacific, large floating plastic debris is tagged and tracked.2 These do however 367 come with the caveat of introducing electrical trash into the environment. For areas close to shore, 368 cheaper, accurate IoT (internet of things) technology can be deployed using conventional 3G 369 networks, or Lora systems to provide better coverage where mobile data is lacking. Iridium 370 satellite connectivity is prohibitively expensive to be deployed into the sea. Accordingly, the use 371 of GPS tags and transmitters to monitor the trajectory of marine litter have been assigned a 372 readiness level of 1. 373
Readiness Level: GPS tags and transmitters
1: Basic research and/or development principles observed and reported
Modeling 374
The abundance of litter in the marine environment has steadily increased over the last few 375 decades and recent studies have showed relatively high concentrations of microplastic particles 376 (particles up to 5 mm) in coastal sediments (Browne et al., 2011). By various means (e.g. transport 377 accidents, inappropriate disposal of packing materials as well as microplastic beads used in 378 cosmetics), different types of plastics enter the water column, with serious ecological implications 379 for marine organisms, such as fatal entanglement in macro plastics or the ingestion of 380 microplastics by fish and birds (Leslie et al., 2011). With these concerns in mind, many questions 381 arise: Which areas probably have the highest concentrations of plastic litter (hotspots)? What are 382 the transport routes of plastic litter in the water bodies and in which areas do they end up? How 383 are different types and sizes of plastic behaving? The fate of the plastics in the marine 384 environment is also uncertain: they might accumulate or degrade due to fragmentation and 385 microbiological decay. To address these issues, an integrated approach is essential that 386 highlights the role of transport and fate models to provide the means to include different processes 387 and investigate their relative contribution. 388
High resolution hydrodynamic models are considered critical to resolving the key marine litter 389 questions, in that they offer a platform that can integrate (and give much greater value to) the very 390 sparsely available observation data (Martinez-Vicente et. al. 2019). An analogous example is the 391 assimilation of the relatively sparse Argo float data into the Mercator global forecast, greatly 392 improving the performance and reliability of the model (e.g. Turpin et. al. 2016). There are 393 certainly technical hurdles e.g. establishing common currency, metrics and uncertainties between 394 specific observation types and models; establishing the necessary sub-mesoscale global nests 395
8https://mcc.jrc.ec.europa.eu/main/dev.py?N=simple&O=394&titre_page=RIMMEL%2520observation%2520Network
9https://www.cls.fr/en/kineis-unique-constellation/
10https://space-science.wwf.de/project/pandasat/
14
of models required for appropriate simulation of litter dispersion and accumulation (D’Asaro et al 396 2018). But the combination of high resolution numerical simulations and sparse observations will 397 certainly play a major role in better understanding global dispersion and accumulation. 398
Numerical modeling of beach litter primarily aims to forecast litter accumulation on beaches to 399 support cleanup efforts and identify potential hot spots (Granado et al., 2019; Haarr et al., 2019; 400 Yoon et al., 2010). One challenge for predicting beach litter accumulation is the fine resolution 401 required, ranging from a few 100 m to 1 km, which can be limiting for forecasting along shorelines 402 that lack high resolution data and oceanographic models (Critchell & Lambrechts, 2016). 403 Combining local and regional high-resolution circulation models with satellite-observed surface 404 debris could provide a basis for forecasts of beaching events. This approach is discussed for 405 forecasting beaching of sargassum and could also be used for marine debris. Research efforts to 406 develop and improve beach litter forecasts through new techniques, such as machine learning 407 and GIS-based tools, are underway but are still in the research and development phase (Aydin & 408 Butler, 2019; Critchell et al., 2015; Critchell & Lambrechts, 2016; Granado et al., 2019; Yoon et 409 al., 2010). 410
A very important auxiliary input for modeling the trajectories of plastics in the ocean are ocean 411 surface currents. The output of regional and global Ocean General Circulation Models (OGCM) 412 can be used to map and predict past and future trajectories of marine plastic. This can assist in 413 identifying sources and accumulation locations (van Sebille et al., 2012). The data used to 414 generate these models include wind speed and direction, mapped sea level anomaly (MSLA), 415 and sea surface temperature, which are available almost daily. These models can be fine-tuned 416 using data from buoys, or GPS tracked plastic pieces (GESAMP, 2019; van der Mheen et al., 417 2019). Three-dimensional modeling of marine plastic is also being developed, for example by the 418 TOPIOS project (http://topios.org/). To improve the knowledge regarding the distribution and 419 possible accumulation zones of marine plastic litter in the north Sea and within the CleanSea 420 project, the microplastics transport is simulated with a hydrodynamic-based paricle tracking 421 model. The model calculates how the position of microplastic particles evolves in time from their 422 release (discharge from rivers such as the Rhine or the Meuse) until the end of the simulation. 423 The settling velocity of the plastic particles in the water system is dependent on the ambient 424 conditions (temperature/salinity) as well as on the particle characteristics (density/size/shape). 425 The developed model is generic and can be extended to other European regional 426 seas.(http://cleansea.eu) (Stuparu et al., 2015). 427
Modeling is a promising approach to improve the existing knowledge regarding the litter dynamics 428 in marine environments and obtain new insights in areas where information is lacking (Thompson 429 et al., 2009). For example, the data regarding the abundance of plastic litter on the seabed is very 430 limited. Also, it is assumed that substantial quantities of plastic litter has accumulated in the 431 natural environment due to the continued input of marine litter over the last decades; however, 432 the location of possible accumulation areas is not well delimited. The modelling approach 433 provides a link between the source and the fate of microplastics. By describing microplastic 434 pathways an overview of estimated accumulation areas is possible, a helpful tool for guided 435 monitoring and data collection campaigns. 436
As modelling of marine litter is still in the research and development phase, we have assigned a 437 readiness level of 1. 438
Readiness Level: Modeling
15
1: Basic research and/or development principles observed and reported
439
16
Table 3. Summary of technology for marine litter monitoring 440
Technology Readiness Level
Size Class
Application Area Pros Cons
Human eye 8
Mega- and macro-litter
• Beach/shoreline litter • Floating/water column
litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Advanced technology not required
• Can be implemented by citizen science volunteers
• Well-developed methods and studies exist
• Depends on regular sampling and commitment of human resources
• Dependent on human error • Resource and time intensive • Requires global agreement
and implementation of comparable methods
Weight 3 Mega-, macro-litter and micro-litter
• Beach/shoreline litter • Floating/water column
litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Allows for relatively quick and simple analysis of beach litter quantities
• Can be linked with voluntary beach clean-up efforts
• Beach litter water content, sand and biofouling can bias results
• Presence of light weight items such as Styrofoam and wrappers may lead to underestimates of beach litter severity
Microscopy 3 Meso- and micro-litter
• Beach/shoreline litter • Floating/water column
litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Provides information about smaller classes of litter
• Provides important information about ingestion
• Sample collection and analysis has risk for contamination
• Sample preparation and analysis varies and requires agreement and implementation of comparable methods
• Time consuming • Human error in identifying
material types
17
Spectroscopy 1 Meso- and micro-litter
• Beach/shoreline litter • Floating/water column
litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Provides information about types of plastics in a sample
• Can provide information about the fate and breakdown of litter
• Time consuming and expensive
• Consistent sample preparation methods not agreed upon
• Limited number of samples can be analysed
Mass Spectronomy
1 Micro-litter • Beach/shoreline litter • Floating/water column
litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Provides information about chemicals associated with litter (such as contaminants)
• Time consuming and expensive
• Limited number of samples can be analysed
Visual Imagery and Video
6 Mega- and Macro-litter
• Beach/shoreline litter • Floating litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Simple and affordable technology
• Variety of systems available including cameras attached to air planes, drones, and submersibles
• Access to hard to reach beaches
• Limited to large debris items • Image processing can be time
consuming
Hyperspectral Imaging
1 Macro-litter • Beach/shoreline litter • Floating litter • Sources of marine litter
• Ability to survey large areas in short periods of time by using satellites, planes or drones
• Access to hard to reach beaches
• Regulatory issues can restrict areas of operation of airborne platforms
• Limited to large debris items • Imagery can be limited by
weather conditions • Image processing can be
challenging and time consuming
• Further research and
18
validation required
Synthetic Aperture Radar
1 Mega-litter • Beach/shoreline litter • Floating litter • Sources of marine litter
• Ability to survey large areas in short periods of time by using satellites, planes or drones
• New sensors and processing tools are in development
• Can be used to identify convergent zones where marine litter accumulation is likely
• Most high resolution data is commercial
• Limited to large debris items • Processing data is resource
intensive
GPS tags and transmitters
1 Mega- and macro litter
• Floating/water column litter
• Can provide information about pathways of marine litter
• Data can improve modelling and source identification efforts
• Iridium satellite connectivity is prohibitively expensive to be deployed into the sea
• Introduces electronic litter into the marine environment
Modelling 1 Mega, Maco-, Meso- and Micro-litter
• Beach/shoreline litter • Floating/water column
litter • Ingestion of marine
litter/entanglement • Sources of marine litter
• Predictive ability can support identification beach litter hot spots in areas lacking on the ground data
• New processing technologies such as machine learning and GIS-based tools show promise
• Many oceanographic models are not at a high enough resolution to predict beach litter
• Running of models can be resource intensive, requiring high levels of computing resources
• Additional research and testing required before models can be used for decision making
441
19
Section 2: Existing marine litter databases and major 442
datasets 443
In order to accurately make recommendations about how to effectively develop a global 444 monitoring platform and inform action for marine litter, it is important to understand what existing 445 platforms and data are available. In the past two decades, there has been a steady increase in 446 the amount of data, reports, and studies related to marine litter. In the past five years alone, the 447 level of information and work focused on marine litter and subsequent areas of interest, has 448 spiked. Through expert input, the areas of interest within marine litter include: marine litter found 449 in the water column, marine litter ingestion and entanglement, marine litter on the 450 seabed/seafloor, marine litter on beaches and shorelines, sources of marine litter, waste 451 management, the plastic life-cycle and microplastics. Figure 1 shows an analysis based on a Web 452 of Science Search that highlights the trends of marine litter research over the past 20 years. 453
454
Figure 1. Research articles published about marine litter and sub categories of marine litter: beach litter, ingestion and entanglement, seabed and seafloor, water column, and sources of marine litter. Data was collected using a Web of Science search and analysis from 1990-2019.
As research and data about marine litter has become more readily available to global audiences, 455 many databases have been created or are being created to house these large repositories of 456
0
1000
2000
3000
4000
5000
6000
7000
Marine Litter Peer-reviewed Publications from 1990-2019
Microplastics
Waste Management
Sources of Marine Litter
Marine Litter on the Beach/Shoreline
Marine Litter on the Seabed/Seafloor
Marine Litter Ingestion and Entanglement
Marine Litter in the Water Column
Marine Litter
20
data, to make them useful to decision makers (managers, policy makers, etc.) and the scientific 457 community. After seeking input from groups, research teams, private, and public sector 458 organizations worldwide, we have compiled an extensive, though not complete, summary of 459 available marine litter databases and datasets. 460
The live results of the inventory can be viewed here. 461
This inventory highlights the specific areas of interest the data covers, the data management 462 protocols of the data, collection and analysis methods, region(s) the data covers, and any other 463 relevant information. 464
Based on primary survey results obtained from 27 databases and datasets, the vast majority of 465 the databases contained information regarding beach and shoreline, followed by seafloor/seabed, 466 water column, sources of marine plastic, and plastic ingested by biota. A number of the beach or 467 shoreline monitoring databases work in conjunction with citizen science groups to collect litter 468 data. Projects and programs from governmental bodies, NGOs, private enterprises, research 469 institutions and universities from across the world are working to solve the marine litter crisis. 470 Listed below are a few examples of marine litter databases ranging in scope, institutional goals 471 and litter area of interest. 472
473
Figure 2. Results of the survey conducted for this paper about marine litter databases and datasets. Participants provided information about their main areas of focus of data collection and analyses were in regards to marine litter.
The NOAA Marine Debris Monitoring and Assessment Project (MDMAP) run by the NOAA 474 Marine Debris Program focuses its efforts on beach litter collection and counting. The database 475 collects their data with the help of citizen science efforts from partner organizations and volunteers 476 conducting shoreline surveys. The database accepts data from shoreline surveys using the NOAA 477 protocol from anywhere in the world, but most of the data is collected from the US Coastal Zone, 478
0 5 10 15 20 25 30
Total
Beaches/Shoreline (Beach litter)
Water Column
Seafloor/Seabed
Plastic Ingested by Biota
Sources of Marine Plastic
Other
Preliminary Marine Litter Database Survey ResultsBased on submissions through August 1, 2019
21
and predominantly the west coast. The MDMAP collects data for debris larger than 2.5 cm in the 479 longest dimension. Within the database both flux accumulation/flux data and 480 standing/concentration data are available, both collected using surveys of specified areas of the 481 shoreline. Data is verified by NOAA staff and reviewed before any of it is published. Anyone can 482 request access to the database, wherein requests are approved by NOAA and then all verified 483 data is reportable/downloadable. 484
485
Figure 3: The Marine Debris Monitoring and Assessment Project Database (MDAMP v.2.0.18, viewed on November 20, 2019).
Marine LitterWatch is operated by the European Environment Agency and focuses on litter on 486 the coastline of most of Europe in order to “strengthen Europe’s knowledge base and provide 487 support to European policy making.” The database includes a total count and itemized breakdown 488 of all items collected/observed. Additionally, the specific locations sampled are broken down by 489 total cleanups, average amount collected per cleanups, and the organization who aided with the 490 cleanups Marine LitterWatch functions as a mobile application for volunteer organizations and 491 Regional Seas programs in Europe to participate in cleanups. The application is used to survey 492 a given area of clean up based on specific items broken into categories of plastic, cloth/textile, 493 and glass/ceramics with sub-sections within those categories. At present, the Marine LitterWatch 494 data represents the effort made by the communities collecting it and is therefore illustrative of the 495 amount and type of items found on the surveyed beaches. Additional handling is required for 496
22
using this data for further statistical purposes. The datasets are also not quality checked or 497 monitored once the data is input into the survey. The EEA wide policy on data management, 498 access, and sharing, is open, free, readily available access to data. 499 https://www.eea.europa.eu/legal/eea-data-policy/data-policy 500
501
Figure 4: The Marine LitterWatch Data Viewer (viewed on November 20, 2019).
The Deep-Sea Debris Database is operated the Japan Agency for Marine-Earth Science and 502 Technology (JAMSTEC) and is a composite of filmed and photographed debris found on the 503 seafloor off the coast of Japan and the Pacific. The images are attributed to the location, date 504 observed, type of debris (plastic, glass, rubber, cloth, etc.), whether organisms were found 505 interacting or near the debris, the characteristic of the sediment, as well as the depth the debris 506 was found. The database has a total quantity of debris observed broken down by type of debris. 507 Cameras are sent below the surface of the water to the seabed and then a specified area is filmed 508 and observed for the debris. The raw data and images are available on the database. Some of 509 the data are protected as intellectual property and labeled as such, but otherwise the data is open 510 and available for others to use. In regards to the ownership of the data, the data and samples 511 collected through the use of JAMSTEC facilities and equipment belong to JAMSTEC. The 512 organization manages the data and samples that can be used by organizations, institutes and 513 researchers for scientific and educational purposes. They promote the use of their data to help 514 industrial and society. Data used by industry may be charged, but all other scientific and 515 educational use of data will be free charge. 516
23
517
Figure 5: The Deep-sea Debris Database (viewed on November 20, 2019)
The Coastal Observation and Seabird Survey Team (COASST) is a part of the University of 518 Washington and focuses on beach litter, ingested litter, as well as sources of this litter in parts of 519 Washington and Oregon, USA. This database provides information on counts and item-specific 520 characteristics (item type, color, material, size, loops, floppiness, brands, logos, languages, 521 shininess, biofouling, weathering, intactness etc.) of items observed during standardized beach 522 surveys following specific protocols for sampling debris between 2.5mm and 2.5cm; 2.5cm and 523 50cm; and greater than 50cm, respectively. To collect this information, citizen scientist volunteers 524 are trained to collect data following the standard protocols developed by COASST. “By 525 collaborating with coastal residents, natural resource management agencies and environmental 526 organizations, COASST works to translate long-term monitoring into effective marine 527 conservation solutions and responsible marine stewardship.” The raw data from collections are 528 available upon request but are not published. The team requires a data use agreement to 529 establish terms of use and the data is quality analyzed and controlled by the team. Additionally 530 post processing procedures are performed to ensure the validity of the data. 531
24
Figure 6: Map showing the locations of data collected for the COASST program as of 2018 (Image credit: Hillary Burgess).
The European Marine Observation and Data Network (EMODnet Chemistry) Marine Litter 532 Database (https://www.emodnet-chemistry.eu/marinelitter) is part of EMODnet Chemistry, one of 533 the seven thematic portals of EMODnet11. EMODnet Chemistry is operated at a European scale 534 through a network of National Oceanographic Data Centers and monitoring agencies coordinated 535 by OGS (Istituto Nazionale di Oceanografia e di Geofisica Sperimentale), an internationally 536 oriented public research institution in Italy. The Marine Litter Database (Molina Jack et al., 2019) 537 offers the first pan-European data on marine litter, namely beach litter (Addamo et al., 2018; 538 European Commission, 2018), floating microlitter in the water column and litter on the seafloor 539
11 www.emodnet.eu
25
from fishing trawlings. Data are assembled from national and regional marine monitoring 540 programmes from across European member states and bordering countries e.g. Ukraine, Russia, 541 Georgia and Montenegro. The database has been developed in strong collaboration with the Joint 542 Research Centre (JRC) of the European Commission to ensure it compiles data requested by the 543 European Commission’s Marine Strategy Framework Directive (MSFD) and by the Regional Sea 544 Conventions. It builds on existing databases, mainly OSPAR/MCS for beach litter in NE Atlantic 545 and ICES DATRAS for sea floor litter in NE Atlantic and Baltic Sea. Additionally, it hosts litter data 546 from wider monitoring and observing programmes, including scientific research, citizen science 547 and specific initiatives like samples collected from racing yachts, in partnership with the Volvo 548 Ocean Race. The database includes data quantities and types of beach litter, sea floor litter and 549 floating microlitter. The online EMODnet Chemistry platform offers a products viewer and access 550 service where marine litter geospatial data can be discovered, viewed and visualized as pan-551 European map layers. Datasets are available for download together with metadata to describe 552 the data collection and acknowledge the original data collector. Generally, there are no restrictions 553 to access or reuse of the raw data available in the database (https://emodnet-554 chemistry.maris.nl/search), and where specific access requirements exist e.g. for particular 555 countries, this is specified in the database. 556
557
Figure 7: The EMODNET Chemistry – Litter viewing and downloading service. Map showing the location of beaches (legenda specify the different reference lists used to describe litter items) and seafloor litter surveys (legenda specify the different sampling gears used during the surveys) (viewed on November 25, 2019).
The Australian Marine Debris Database is organized by an Australian NGO called the Tangaroa 558 Blue Foundation as part of the Australian Marine Debris Initiative. This database collects 559 information on beach litter primarily on the Australian Coast with some data also available from 560 the Asia Pacific and Oceania region. The litter is categorized by debris items with 140 different 561 categories. Categories are based on material type and item name. To collect data volunteers 562 perform beach clean-ups of coastal areas both on land (beaches) and sea (near-shore surface 563 levels). Litter is counted and itemized based on the specifications. Approximate weight of litter 564 and length of area of cleanup is reported. Photos may also be attached. Data collected is vetted 565
26
before approval. The database has an open access policy that allows for a specific set of data 566 reports to be generated for community groups, schools and partner organisations to assist in 567 identifying marine debris trends and creating local source reduction plans. When any data is used 568 publicly for any purpose both the Australian Marine Debris Initiative and the data contributor must 569 be acknowledged. This information is available by emailing [email protected] with both the 570 location and date of the data being requested. Additionally, there is a data management system 571 in place - data submitted is placed in holding folder for vetting before being accepted into the 572 database. 573
574
Figure 8: The Australian Marine Debris Initiative Database (viewed on December 3, 2019)
The TIDES (Trash information and Data for Education and Solutions) database is operated by 575 the Ocean Conservancy and focuses on cataloguing and collecting litter found on beaches, 576 shorelines and in the water column. This database contains information on the total mass of trash 577 collected, the total number of trash bags filled, the total distance of area covered. An itemized list 578 of total trash collected broken down by most likely to find items, fishing gear, packaging materials, 579 and other items such as personal hygiene products, smaller trash items (less than 2.5 cm) and 580 items of local concern. Data can be collected and reported as land based trash, underwater, or 581 trash collected by watercraft. The main form of collection of information comes from annual 582 international cleanup events and a mobile app called Clean Swell are used to collect and itemize 583 trash found near and in bodies of water. Groups or individuals collect trash and tally the total 584 number of specific items found as well as the overall mass of the total trash. The data is recorded 585 to the TIDES database and publicly available, though there is no specific database/dataset 586 management protocol, but site specific datasets are available and archived from past years data 587 collected. 588
27
589
Figure 9: The Trash information and Data for Education and Solutions Database (viewed on December 3, 2019)
Litterbase is a global portal organized by the Alfred-Wegener-Institut Helmholtz-Zentrum für 590 Polar and Meeresforschung. This portal has information on litter present on beaches/shorelines, 591 the water column, the seafloor, ingested plastics scoping across oceans, rivers, lakes, and other 592 inland waters. In the portal, information can be found regarding: quantitative geo-referenced data 593 on aquatic and terrestrial debris, microplastics and nanoplastics from the peer-reviewed literature, 594 quantitative geo-referenced data on effects of marine debris, microplastics and nanoplastics on 595 aquatic and terrestrial biota from the peer-reviewed literature (field studies) and reports of impacts 596 of marine debris, microplastics and nanoplastics on aquatic and terrestrial biota from the peer-597 reviewed literature (laboratory studies, species list). 598
599
28
Figure 10: Global map of litter distribution in Litterbase (Image source: https://www.maritime-executive.com/article/ocean-litter-portal-established)
The Global Ghost Gear Initiative: Data Portal is organized by the Global Ghost Gear Initiative 600 of the Ocean Conservancy. This initiative works to find fishing gear that has been lost, abandoned, 601 or otherwise discarded. This initiative works with global partners including the fishing industry, 602 private sector, NGOs, academia, and governing bodies. The data portal has data from the US 603 coasts, the European Coasts and the Asia-Oceania Pacific region. In the data portal, information 604 about different types of “ghost gear" is available including found nets, lines, pots and traps. Total 605 counts and location, dates, gear class. The data is collected using volunteers and partners that 606 upload data to their mobile application “GGGI Ghost Gear Reporter.” Bulk upload is available on 607 their website as well. Additionally, all data is available on the data portal, and specific 608 measurements are available upon request as well. 609
610
Figure 11: The Global Ghost Gear Initiative Data Portal (viewed on December 3, 2019).
611
612
29
Section 3: Indicators and Applications of Technologies 613
The primary questions to be addressed by marine litter monitoring are: 614
1. What is the abundance, distribution and composition of marine litter, and are these 615 attributes changing over time? 616 617
2. What are the main sources of marine litter, and are they changing over time? 618 619
3. What are the impacts of marine litter, and are they changing over time (Ryan et al., 2009)? 620 621
4. What are the principal dispersion and accumulation pathways and what are their 622 variability? 623
In this section, we review existing and developing indicators to address these questions as well 624 as proposed indicators for reporting on SDG 14. 625
Indicators for abundance, distribution and composition of marine litter 626
Indicators for the abundance, distribution and composition of marine litter have been developed 627 for beach/shoreline litter, floating/water column litter and seafloor litter (GESAMP, 2019; UN 628 Environment, 2018). These indicators seek to provide a measure of the state of marine litter in 629 the environment (GESAMP, 2019). 630
Beach/shoreline litter 631
Indicators on beach litter are produced using various methods that take into consideration types, 632 quantities, distribution and fluxes. Some studies record numbers of various types of marine litter 633 while others look at the mass of litter with some studies looking at both (Galgani et al., 2015; 634 Galgani, Hanke, et al., 2013). Beach litter indicators can be used to focus mitigation measures 635 and evaluate the effectiveness of legislation and regulations by providing information on the 636 amounts, trends and sources of marine litter (OSPAR, 2010). Beach litter indicators are the most 637 developed and common indicators and have been used extensively for analysis in regions 638 including the North-East Atlantic, Baltic Sea and United States (European Commission JRC, 639 2013; Hardesty et al., 2017 ; OSPAR, 2017). 640
Focusing on distribution, the GESAMP report highlights the importance of understanding the 641 physiology of a shoreline. The dynamic nature of shorelines, due to both oceanographic and 642 meteorological factors, such as tides, waves and currents, and winds and rain, are dominant in 643 determining how marine litter will end up on beaches. Additionally, the nature of the shore line, 644 specifically the surface structure and slope, will determine what type of litter remains on the beach 645 and where that litter is located over space and time. Specifically, they highlighted Ekman 646 transport, a process in which on and offshore winds will blow floating litter onto or off of the 647 shoreline which causes pronounced currents both on and offshore, as a means of understanding 648 the flux process between floating and shoreline litter. 649
Additionally, tourism and increased human activity is a good indicator of beach litter quantities. 650 Intuitively, seasonal increases of visitors to the beach will lead to increased quantities and types 651
30
of litter load to an area. Conversely, high levels of human activity can also indicate lower levels of 652 specific larger types of litter due to organized beach sweeps (Opfer et al., 2012; Ryan et al., 2009). 653 Using temporal, geographical, and oceanographic metrics for indicators of when beach litter will 654 in high quantities and when certain types of litter will be present, is an effective way to know when 655 monitoring activities should be conducted. 656
Floating and water column 657
The distribution of marine litter on the surface and throughout the water column is influenced by 658 a variety of factors including ocean circulation, material density, degradation and biofouling. The 659 composition of marine litter in the water column ranges from large items such as abandoned, lost, 660 discarded fishing gear (ALDFG) to microplastics (GESAMP, 2019). Indicators for floating and 661 water column debris are essential for knowing what sampling strategy to adapt when monitoring 662 marine debris. In a sort of circuitous route, the indicator needed to identify marine litter quantity 663 and distribution floating on the surface or in the water column, depends on what type of material 664 you are trying to monitor or sample. 665
Generally, there are a few primary indicators that can be used to determine where marine litter 666 will be in the open waters, how much will be there, and the type. As with beach or shoreline litter, 667 temporal variations play a big factor in indicating the location and distribution of litter, including 668 tidal conditions, short term wind and rain events, and seasonal extremes or anomalies. There are 669 also specific types of litter or plastic that are more or less dense than water and so depending on 670 the structure, make up and size of the litter, the distribution and composition of litter within the 671 water column is slightly more straightforward to determine; for example, polystyrene will sink while 672 polyethylene and polypropylene polymers would be expected to float (GESAMP, 2016). 673 Understanding shipping and fishing pathways and monitoring may also be a useful indicator as 674 to where marine debris may be found, especially debris in the water column. Most of the debris 675 found in the ocean is land based, but there is still a significant portion of debris entering the water 676 from sources at sea (NOWPAP CEARAC, 2007). Very high-resolution satellite images, UAV data 677 and ship mounted cameras can be used to indicate mega litter conglomerates on the sea surface. 678
Seafloor 679
From shallow areas near reefs, to deep trenches, litter has been found all across the seafloor. 680 Indicators for monitoring sea floor litter are a more nuanced than indicators for beach litter or 681 floating and water column litter, because the sea floor is a sink for marine litter (Galgani et al. 682 2000, Pham et al. 2014, Woodall et al. 2014, GESAMP 2019). There are a few factors that can 683 be used to indicate both type and distribution of litter on the seafloor, but there are two dominant 684 non-naturally occurring indicators of potential seafloor litter. The first is proximity of maritime 685 activities, such as fisheries, aquaculture, shipping, construction, energy extraction and 686 recreational activities (Pham et al. 2013, Loulad et al. 2017). The second is shore based leakage 687 or run off points, like major river deltas, populated and industrialized coastlines and coastal 688 tourism. Though not all seafloor litter is macro in scale (there is a significant issue of microplastics 689 in sediments), a key indicator of type of litter found on the seafloor is the physical characteristics 690 of litter, especially density and size. 691
While time and seasonal trends are hard to use as indicators, especially in regard to quantity, due 692 to the lack of baseline studies and observations of seafloor litter, environmental factors of the 693
31
seafloor may be a key indication of the possible presence of litter. Water depth, seafloor 694 topography, surface and deep water currents may be an indicator of distribution (GESAMP 2019). 695
Indicators for sources of marine litter 696
Using information about spatial distribution of floating litter, with current, tidal, and riverine 697 information can be a useful source indicator for marine debris. This may allow for important 698 evidence about the pathway and input zone, which can be used to determine the potency of the 699 source as well as the efficacy of any management practices in place. The use of specific items 700 as indicators of sources or pathway of marine input is a useful practice, such as items from 701 industrial or fishing vessels. 702
Land-based sources (rivers, coastal disasters) and sinks (beaches) could be accounted for. The 703 major land-based sources of marine plastic include land-fills, floodwaters, industrial outfalls, 704 discharge from storm water drains, untreated municipal sewerage, and littering of beaches and 705 coastal areas from tourism and other activities. Information on these sources could be harvested 706 from existing databases, social media and public documents. 707
The integrated information and model system would provide a basis for risk assessments. For 708 example, a candidate for assessing the risk of seafood contamination from ocean plastics is the 709 functional dependency network analysis (Pinto & Garvey, 2013), which this model system would 710 support. Likewise, the model system would facilitate cost-benefit analyses for mitigation means. 711
This integrated system also would allow for a scenario-based exploration of possible futures. After 712 careful validation and calibration, this model could be used to assess future trajectories for ocean 713 plastics based on scenarios of plastic production, waste management, recycling and reuse 714 practices, as well as efforts to remove plastics from the ocean. Desirable futures could then be 715 connected back to transformative policies that need to be implemented to ensure such futures. 716
Plastic debris in rivers, including the mouths of rivers and estuaries 717
Main sources of marine litter entering the ocean through rivers are due to improperly managed 718 plastic waste, including failed recycling, inadequate sewage systems, and inadequate disposal 719 (Jambeck et al., 2015). A combination of an intensive 2-week in situ sampling program with 720 hydrological data showed that the Saigon river, Vietnam, carried macroplastic loads at least four 721 times higher than previously estimated (Van Emmerik, 2018). This underlines the importance of 722 case studies in those rivers that knowingly contribute significantly to the flow of plastic into the 723 ocean. The Ocean Cleanup initiative (https://theoceancleanup.com/rivers/) is working with 724 governments to prevent plastic from entering the world’s oceans from rivers from 1000 of the most 725 polluting rivers, all over the world, by 2025. 726
Sediment outflows at river mouths, indicative and correlated with land-based sources of pollution 727 might be a potential indicator for plastic debris. Sediment samples in estuaries could also provide 728 information on plastic contents, potentially given time variability over the last five to seven 729 decades. 730
In addition to estimates of plastics at river mouths and in estuaries, it would be important to map 731 the input of plastic into the rivers. Variables such as watershed population, sources of waste and 732
32
leakages into the environment, management practices, and runoff would be important auxiliary 733 data to harvest from existing sources. 734
Marine litter debris from ocean activities (shipping, fishing, mining) 735
Many sea-based activities contribute to marine debris. Important contributions come from fishing 736 and aquaculture, shipping (e.g., transport, tourism), dredged material, offshore mining and 737 extraction, sewage sludge and illegal dumping at sea such as waste streams containing plastics 738 and microplastic. As most sea-based sources of plastic come from ship presence or traffic, the 739 comprehensive available Automatic Identification System (AIS) data provides a database of 740 valuable information about ships and their movements. While various free sources of AIS data 741 exist online, these are limited in scope. The full database is available for purchase. Based on this 742 full database, pattern recognition and matching algorithms could be used to match hotspots of 743 marine litter with ship presence, taking into account the trajectories of these hotspots based on 744 ocean currents. This would allow determination of ship size, type, and flag country to identify the 745 most likely polluters. 746
Knowing where the most important fishing areas are at any given one point in time (e.g., using 747 the such as with Global Fishing Watch, https://globalfishingwatch.org) would help to detect major 748 potential sources and locations of ghost gear. 749
A key convention for the International Maritime Organisation (IMO) is the International Convention 750 for the Safety of Life at Sea (SOLAS). Regulation 19 of Chapter V of SOLAS - carriage 751 requirements for shipborne navigational systems and equipment – lists the navigational 752 equipment to be carried on-board ships in accordance to ship type. All ships are required to carry 753 AIS which must be able to provide information about the ship to other ships and to coastal 754 authorities automatically. More specifically, regulation 19 of SOLAS Chapter V requires AIS to be 755 installed on-board all ships which are of 300 gross tonnage and upwards engaged on international 756 voyages, cargo ships of 500 gross tonnage and upwards not engaged on international voyages 757 and all passenger ships irrespective of size. 758
This means that many fishing vessels cannot necessarily be tracked through AIS. However, it is 759 mandatory for all fishing vessels engaged in commercial activities to broadcast their positions via 760 encrypted satellite communication every 2 hours. This system, known as Vessel Management 761 System (VMS) was specifically created to monitor national fishing fleets and foreign vessels 762 fishing within national waters is available only to national governments authorities and those they 763 share it with. By engaging national authorities as well as FAO, VMS data at global level can be 764 used to track, monitor, model and evaluate the sources and locations of fisheries ghost gears. 765
Aquaculture is also a known source of lost fishing gear and apparatus. Locations of these 766 activities can be detected reliably with high-resolution imagery (Trujilo et al., 2012). 767
The Marine Environment Protection Committee (MEPC) of the International Maritime 768 Organisation (IMO) has agreed an Action Plan (IMO, 2018) to address Marine Litter from ships 769 (including from fishing vessels). Building on the existing policy and regulatory frameworks such 770 as the MARPOL Convention (MARPOL, 1973) and the Convention on the Prevention of Marine 771 Pollution by Dumping of Wastes and Other Matter (the London Convention and Protocol for short, 772 LCP, 1972), the action plan introduces new supporting measures to address the issue of marine 773 litter from ships. The Global Platform can feed into the Action Plan and respond to the 774 requirements of this Action Plan. 775
33
Marine litter from coastal disasters 776
The modern built environment includes a large fraction of plastic material. In 2015, 72 million tons 777 of plastic went into building and construction (with an average use time of 35 years)12 [13]. 778 Considering the migration of the global population, a large fraction of this is located in the coastal 779 zone or in flood zones and thus exposed to hydro-meteorological hazards. This rapidly increasing 780 exposure of the built environment to floods and storms has increased the likelihood of plastic and 781 other debris entering the ocean. The risk is further exacerbated by the likely increase of the 782 frequency and intensity of hydro-meteorological hazards due to modern climate change. 783
Information on the amount of marine debris resulting from coastal disasters is urgently needed. 784 Databases compiled by insurances, real-estate companies and municipalities could be harvested 785 to estimate and map the plastic integrated the built environment. Overlaying this information with 786 disaster assessment would provide a basis to quantify the amount of plastic and other debris 787 washed into the ocean during major hazardous events. 788
Primary Microplastics 789
Microplastics in the environment are categorized as primary and secondary microplastics. 790 Boucher and Damien (2017) define primary and secondary microplastics as follows: 791
• Primary microplastics are plastics directly released into the environment in the form of 792 small particulates. They can be a voluntary addition to products such as scrubbing agents 793 in toiletries and cosmetics (e.g. shower gels). They can also originate from the abrasion 794 of large plastic objects during manufacturing, use or maintenance such as the erosion of 795 tyres when driving or of the abrasion of synthetic textiles during washing. 796 797
• Secondary microplastics are microplastics originating from the degradation of larger 798 plastic items into smaller plastic fragments once exposed to marine environment. This 799 happens through photodegradation and other weathering processes of mismanaged 800 waste such as discarded plastic bags or from unintentional losses such as fishing nets. 801
Emission estimates of primary microplastics can be used as indicators of marine miroplastic 802 pollution. Primary microplastics include tire dust/particles, road markings, synthetic textiles, 803 maritime coatings, personal care products, plastic pellets and artificial turf (Boucher & Damien, 804 2017; Wang et al., 2019). Emission estimates can then be linked into estimates of micoplastics 805 entering the aquatic environment through various pathways (e.g. domestic sewage, road runoff, 806 wind, adjacent waters) (Burton, 2017; Lassen et al., 2015; Verschoor et al., 2016). 807
12 Parker, L., 2018. We made plastic. We depend on it. We’re drowning in it. National Geographic, June 2018 issue.
34
For example, a recent study by Wang et al. (2019) utilized this process to estimate the 808 contributions of various items to primary microplastic emissions and estimate the amounts 809 entering aquatic environments in mainland China (figure X). 810
811
812
Figure 12: Contribution of various sources to total primary microplastic emissions (a) and the amounts entering the aquatic environment (b) in mainland China in 2015. Source: Wang et al., 2019.
These estimates can be validated by the analysis of microplastics in known sources, such as 813 storm water, and the marine environment (Sutton et al., 2019) though additional analysis and 814 research is needed before primary microplastic emissions can be routinely used as indicators. 815
Waste Management 816
Effective indicators for waste management are an important measure on the impact waste creates 817 on the marine environment. Waste management indicators look at production of solid waste, 818 sewage treatment, tourist activities (ASEAN, 2007; Prabhakaran, 2013; Tanguay et al., 2012; 819 WTO, 2003). A recent study by Prabhakaran (2013) summarized indicators related to marine 820 waste management (Table 4). 821
Table 4: Marine waste management indicators (Prabhakaran, 2013)
Measurement Indicator Nature and type of solid waste Level and time of degradation based on
type of waste Number of waste disposal bins Sufficient capacity for collecting solid
waste Community involvement in solid waste management such as waste separation, waste reuse and recycling
Percentage of community involved in solid waste management
Numbers of tourists who intend to return
Tourist satisfaction of the site
35
Promotion of environmental activities in order to encourage involvement of hotel staff, clients and suppliers to participate in environmental practices
Level of stakeholder’s awareness in environmental activities
Cost of clean-up Cost for coastal contamination Marine habitat and species changes
Impact on marine species
Percentage of local population who agree that suitable enforcement is in place (local questionnaire)
Development control and enforce
Percentage (quantity) of revenue (key use areas) dedicated to protection
Site protection and funding for protection
Number of reported pollution or contamination events per annum (by month) in watercourses receiving effluents
Effect of sewage treatment
Quantity of waste collected from public areas and beach
Maintaining clean image for the destinations
Material Life Cycle 822
The life cycle of a material consists of: 823 1. Extraction of raw materials, 824 2. Manufacturing and distribution, 825 3. Use and/or consumption, 826 4. Re-use, 827 5. Recycling of materials, 828 6. Energy recovery, and 829 7. Disposal (JRC, 2012). 830 831 Indicators for the various life cycle stages of plastics (and other components of litter) can provide 832 important information about the current and projected status of marine litter. These indicators can 833 also provide information to industry and policy makers regarding material selection and regulation. 834 For example, comparison between plastic and aluminum containers in the United States from 835 1960 – 2017 demonstrates a contrast between the life cycles of plastic and steel containers in 836 terms of production, recycling and disposal (Figure 13). 837 838 839
36
Figure 13: Plastic and Steel Containers and Packaging Waste Management in the Untied States from 1960 – 201713.
840
Indicators specifically related to marine litter and material life cycles include analysis of 841 improperly disposed litter that subsequently leaks into the marine environment as well as 842 indicators for the residence time of litter in the marine environment. 843
Indicators for impacts of marine litter 844
Biological Impacts 845
Marine organisms regularly interact with litter deposited into the ocean. Whether there are filter 846 feeders incidentally consuming microplastics, birds nesting on floating debris on the ocean 847 surface or beach, larger fish eating litter that travels upwards through the different trophic levels, 848 or coral reef habitats being disturbed by litter on the ocean floor, litter and debris impact myriad 849 marine biota through a number of different means. Monitoring how, where, why, and when 850 organisms interact with litter is crucial for the safety and wellbeing of the oceans. 851
Biological and classical indicators play a pivotal role in the monitoring of marine litter. Because 852 not all litter will be collected or counted, the biological indicators act as a way to measure the 853 impact marine debris is having on the environment and can also serve as a way to assess the 854 impact of a specific measure or policy set in place. For example, the INDICIT II project has found 855 an effective biological indicator should be “accurate, sensitive, reliable and easy to use for all the 856 stakeholders in order to be applied to a large geographic area.” Sea turtles, crustaceans and fish 857 are useful indicators because they tend to ingest of become entangled in marine debris, have a 858 large spatial distribution, and use all ecological marine components from the seabed to sea 859 surface14. 860
13 https://www.epa.gov/facts-and-figures-about-materials-waste-and-recycling/containers-and-packaging-product-specific-data#PlasticC&P
14 https://indicit-europa.eu/cms/wp-content/uploads/2018/09/Protocole_v7_hd.pdf
37
Entanglement and ingestion are two biological indicators that will provide information about the 861 interactions between organisms and marine debris. Monitoring entangled organisms can indicate 862 changes in the abundance of debris responsible for entanglements7. Entanglement also serves 863 an indicator for the harm caused by the incorporation of marine debris into nests of breeding 864 birds6. While it occurs less frequently than ingestion, using a consistent monitoring approach could 865 potentially allow for entanglement to be used as an indicator for the success of mitigation efforts.15 866 In addition, the presence of plastic items in nests can be used as an indicator of the amount of 867 litter in the natural environment near their nesting areas as well as a risk for entanglement6. 868
Ingestion of marine debris can be useful as an indicator for a number of different things. The first 869 is that, plastic content of a bird stomach can be used to determine regional differences in the 870 abundance of marine debris. Comparing plastic loads of birds in different regions can act as a 871 way to show where more pollution is acting as both a source and a sink7. Using stomach contents 872 of fulmars, the OSPAR Commission developed an indicator to demonstrate the changes in 873 quantities of floating debris in the North Sea as well as the impact it has on biota3. 874
The CleanSea Project, developed a series of considerations to make when selecting or 875 implementing an organism as a bio-indicator (as opposed to collecting samples from naturally 876 available species) that would aid in both consistency and accuracy. As with experiments, selection 877 should be based on a site by site, case by case basis, but the general guidelines provided for 878 selection are as follows: region specific indicator species; non-threatened or protected species; 879 species that can be kept in cases for easy field deployment or retrieval (such as bi-valves); 880 invertebrate species (require less training and handling than vertebrate species); perform 881 sampling in a cost-effective manner by synergies with pre-existing programs; species which when 882 measure are directly linked to impact and effects (more difficult to achieve); species that are 883 directly linked to measure and could be used to evaluate progress towards targets and 884 effectiveness of mitigation activities (Ryan et al., 2009). The use of these guidelines for selection 885 of a bio-indicator, while not comprehensive and fully incorporating all debris, can be a useful 886 starting point. 887
A more general approach to indicator selection of sentinel species as indicators was developed 888 by Fossi, et al. This study surveyed reports of marine species impacted by debris in the 889 Mediterranean Sea, specifically species that had ingested debris. Based on their findings, they 890 determined 6 key criteria to consider when selecting an indicator based on ecological and 891 biological data. The first was background information, including biological and ecological 892 characteristics of the species as well as the knowledge of the non-affected species to be able to 893 generate a point of comparison. The second was habitat information of the species, including both 894 the habitat and home range of the species (sessile, motile, depth, travel, migration). The third was 895 trophic information and feeding behavior, specifically the feeding mechanic and behavior 896 knowledge so that a wide range of levels of the food scale can be selected. The fourth was spatial 897 distribution of species, which was deemed important because of the spread of debris both across 898 the surface and seabed and throughout the water column. The fifth was commercial importance 899 and conservation status, which may allow for a measure of potential transfer of plastic from sea 900 food to humans. Additionally, it is important to be able to continuously monitor species of concern 901 and how they are impacted by marine litter. The sixth recommendation was to note the 902 documented ingestion of marine litter based on the statistics and data available (Fossi et al., 903
15https://mcc.jrc.ec.europa.eu/documents/201702074014.pdf
38
2018). As with the CleanSea Project, the criteria of selection represent a basis of guidelines and 904 not all the information needed will be readily available and not all species used will be the all 905 telling indicator. With all data collection though, having a consistent basis of how to sample and 906 prepare data will have a number of long term benefits, namely the ability to accurately compare 907 data from across different regions of the world. 908
Indicators for SDG reporting 909
The agreed indicators for reporting on marine plastic litter under SDG Target 14.1.1b are based 910 on the the existing internationally agreed GESAMP guidelines. Sub-indicators beach litter, floating 911 plastic and plastic in the sea column, plastic on the sea floor and additional option indicators 912 included in the approved methodology (table 5, UN Environment 2019). Indicators are categorized 913 into three levels: 914
Level 1: Global indicators 915
• Plastic patches greater than 10 meters (for Areas Beyond National Jurisdiction or 916 Total Oceans) 917
• Beach litter originating from national land-based sources 918
Level 2: National indicators 919
• Beach litter count per km2 of coastline (surveys and citizen science data) 920 • Floating plastic debris density (visual observation, manta trawls) 921 • Water column plastic density (demersal trawls) 922 • Seafloor litter density (benthic trawls (e.g. fish survey trawls), divers, video/camera 923
tows, submersibles, remotely operated vehicles) 924
Level 3: Supplementary indicators 925
• Beach litter microplastics (beach samples) • Floating microplastics (manta trawls, e.g. Continuous Plankton Recorder) • Water column microplastics (demersal plankton trawls) • Seafloor litter microplastics (sediment samples) • Plastic ingestion by biota (e.g. birds, turtles, fish) • Plastic litter in nests • Entanglement (e.g. marine mammals, birds) • Plastic pollution potential (based on the use and landfilling of plastics) • River litter • Other parameters related to plastic consumption and recycling • Health indicators (human health and ecosystem health)
Table 5: Monitoring parameters for marine plastic litter to track progress against SDG Target 14.1 (UN Environment, 2019).
Monitoring parameters (and methods) Level 1 Level 2 Level 3
39
Plastic patches greater than 10 meters* X
Beach litter originating from national land-based sources X
Beach litter (beach surveys) X
Floating plastics (visual observation, manta trawls) X
Water column plastics (demersal trawls) X
Seafloor litter (benthic trawls (e.g. fish survey trawls), divers, video/camera tows, submersibles, remotely operated vehicles)
X
Beach litter microplastics (beach samples) X
Floating microplastics (manta trawls, e.g. Continuous Plankton Recorder)
X
Water column microplastics (demersal plankton trawls) X
Seafloor litter microplastics (sediment samples) X
Plastic ingestion by biota (e.g. birds, turtles, fish) X
Plastic litter in nests X
Entanglement (e.g. marine mammals, birds) X
Plastic pollution potential (based on the use and landfilling of plastics)
X
River litter X
Other parameters related to plastic consumption and recycling X
Health indicators (human health and ecosystem health) X
* This indicator is most useful for areas beyond national jurisdiction or total ocean area, not for 926 national monitoring. 927
These indicators are marked as levels 1, 2 or 3, level 1 being global data or globally modelled, 928 level 2 including national monitoring and level 3 describing supplementary/recommended 929 indicators.930
40
931
Section 4: Existing and Developing Global Data Platforms 932
One challenge regarding ocean data is the proliferation of databases and portals. Recognizing 933 this challenge, several global efforts are in development to create aggregated platforms that 934 search and/or harvest data from multiple databases and repositories. This section provides 935 summaries of existing and developing platforms that could host or be leveraged (e.g. support a 936 portion of the platform or supply data) to a global marine litter platform. 937
Global Earth Observation System of System (GEOSS) Platform 938
The Group on Earth Observations (GEO) is a partnership of more than 100 national governments 939 and in excess of 100 Participating Organizations that envisions “a future wherein decisions and 940 actions for the benefit of humankind are informed by coordinated, comprehensive and sustained 941 Earth observations” (GEO, 2005). The GEO community is creating a Global Earth Observation 942 System of Systems (GEOSS) to better integrate observing systems and share data by connecting 943 existing infrastructures using common standards.16 944
The GEOSS Platform17 proactively links existing and planned observing systems around the 945 world and supports the development of new systems where there are gaps. The GEOSS Platform 946 promotes the use of common technical standards so that data from thousands of different 947 instruments can be combined into coherent data sets. The GEOSS Platform (Figure 3) is a 948 brokering infrastructure. The GEO Discovery and Access Broker (GEO DAB) is the primary 949 mechanism by which all data and information is discovered and accessed. The GEO DAB 950 implements the necessary mediation and harmonization services through Application Program 951 Interfaces (APIs). These APIs allow data providers to share resources without having to make 952 major changes to their technology or standards. 953
Presently, the GEOSS Platform brokers more than 150 autonomous data catalogs and 954 information systems, useful for the different GEO Societal Benefit Areas including data from: 955 CAFF, Data.gov, Data.uk, EEA, GBIF, Iris, JRC Open Data catalog, NASA, NCAR, NOAA, OCHA 956 HDX, RCMRD, UNEP, UNOSAT, USGS, Web Energy Services, WMO WIS, Esri Living Atlas of 957 the World, and many more. Data providers are constantly being added and brokered, according 958 to user needs, and it would be possible to add and broker marine litter data from a variety of 959 sources. 960
The GEOSS Platform is testing evolved capabilities implementing a series of scenarios that illustrate 961 its potential to support access and use of Data and Knowledge as a possible contribution to the 962 implementation of a results-oriented GEOSS. It shows how the GEOSS Platform could potentially 963 provide value to different categories of users, including Earth scientists and policy makers, in finding, 964
16http://www.earthobservations.org/geo_community.php
17http://www.earthobservations.org/gci.php
41
producing and analysing information, ultimately supporting the process of knowledge acquisition by 965 the final consumers. 966
The demonstrated GEOSS Platform capabilities enable: 967 • Harmonized discovery and access of data, information and knowledge from heterogeneous 968
distributed sources 969 • Analytical comparison of resources 970 • Value added products generation and sharing 971 • Knowledge building through mediated collaboration (registration of models, algorithms, data; 972
curation of relations between data, services and publications). 973
974
Figure 14: Components of the GEOSS Platform
The GEOSS Portal currently offers a single access point for searching for and identifying available 975 data sources. The GEOSS Infrastructure is planned to evolve through 2020 – 2022 to include the 976 development of GEO Community Portals, or “hubs” (DeLoatch, 2019). The development of a GEO 977 Marine Litter Community Portal is a possibility. 978
Ocean Data Information System (ODIS) 979
42
The programme "International Oceanographic Data and Information Exchange" (IODE)18 of the 980 "Intergovernmental Oceanographic Commission" (IOC) of UNESCO serves to enhance marine 981 research, exploitation and development, by facilitating the exchange of oceanographic data and 982 information between participating Member States, and by meeting the needs of users for data 983 and information products. 984
IODE currently supports several marine data and information products and repositories including 985 the Ocean Biogeographic Information System (OBIS)19 (a data system for biodiversity and 986 biogeographic data and information on marine life and the Ocean Data Portal20 (a data system 987 that collects, integrates and manages physico-chemical data). IODE also has a number of 988 distributed National Oceanographic Data Centres (NODCs) and Associate Data Unites (AUDs) 989 that work to support data and information management in member states21. 990
Currently IOC data is not accessible through a single portal or platform leading to the 991 recommendation from a 2016 external audit of IOC and its activities for IODE to implement a 992 universal marine data and information system. In response to audit recommendation, IODE 993 produced a concept paper for the development of an Ocean Data and Information System (ODIS) 994 that would improve the accessibility and interoperability of existing data and information linked to 995 and not linked to the IOC. A concept paper outlines a conceptual architecture for the system 996 (Figure 15), as well as, an implementation plan and a Cost Benefit Analysis (Spears et al., 2017). 997
18https://www.iode.org/
19http://www.iobis.org/
20http://www.oceandataportal.org/
21https://www.iode.org/index.php?option=com_content&view=article&id=61&Itemid=100057
43
998
Figure 15: The Conceptual Architecture of ODIS (Spears et al., 2017)
At the thirtieth session of the IOC-UNESCO assembly, IODE was invited to prepare a fully detailed 999 and costed project proposal for ODIS for submission to the IOC Executive Council at its 53rd 1000 session in 2020 (IOC-UNESCO, 2019). If the concept is funded and moves forward, it is possible 1001 that ODIS could support access to global marine litter data. 1002
44
Ocean Data Platform 1003
The Ocean Data Foundation22 is a not-for-profit foundation funded by the Resources Group, the 1004 philanthropic foundation of Norwegian business-man Kjell Inge Røkke. The Ocean Data Platform, 1005 an initiative of the ODF, is an open and collaborative data platform that harnesses the power of 1006 data liberation and data contextualization for the public, industry, academia, science, 1007 policymakers and governments. The platform strives to connect data, people & technology to 1008 drive sustainable ocean governance and blue economy. 1009
The platform is planned to function as an open collaboration with existing data providers and 1010 knowledge hubs for ocean data23. The ODP currently is developing selected key use cases 1011 around which it will build the platform. Depending on the development timeline and functionality 1012 of the Ocean Data Platform, it is possible that this platform could support access to global marine 1013 litter data. 1014
European Marine Data and Observation Network (EMODnet) 1015
EMODnet is a long-term flagship initiative of the EU (funded by the European Commission 1016 Directorate-General for Maritime Affairs and Fisheries). EMODnet has mandate and goal to 1017 deliver open access to aggregated and standardized marine data and data products across seven 1018 thematic areas, namely bathymetry, biology, chemistry, geology, human activities, physics and 1019 seabed habitats. It offers a broad range and wealth of in situ data, in addition to combined data 1020 products with satellite-derived data (e.g. bathymetry). Products range from a Digital Terrain Model 1021 for high-resolution Bathymetry to Seabed Habitat Maps (following EUNIS classification), Vessel 1022 Density Maps (monthly composites) and the marine litter maps of the Marine Litter Database. 1023 EMODnet delivers this in collaboration with other key marine data initiatives including the 1024 Copernicus Space programme (and Copernicus Marine Service) and the Data Collection 1025 Framework (fisheries). EMODnet has an increasingly international user community. In addition, 1026 through its Data Ingestion service and international collaborations EMODnet increasingly offers a 1027 wider coverage of datasets, including beyond Europe. EMODnet Chemistry is one of the seven 1028 thematic portals of EMODnet and provides access to a broad range of chemical data spanning 1029 chlorophyll to dissolved gases and pollutants, including marine litter. Data products are also 1030 available for eutrophication, contaminants and marine litter across 6 European sea and bordering 1031 ocean regions. For marine litter, data are assembled, standardized and aggregated from multiple 1032 gear types and the collected litter data follows the data policy defined by data originators. For 1033 restricted data, a negotiation process between the user and the data originator is facilitated by 1034 the relevant National Oceanographic Data Center. When data are used, acknowledgement of the 1035 data source is requested. EMODnet Chemistry provides access to the litter datasets through a 1036 dedicated discovery and access service (https://emodnet-chemistry.maris.nl/search) allowing to 1037 search by the available parameters (space, time, matrix, group of variables, discovery parameter, 1038 data distributor and country). In addition, the aggregated datasets are described in the product 1039 catalogue (https://www.emodnet-chemistry.eu/products/catalogue) providing also information on 1040
22 www.oceandata.earth
23 https://www.revocean.org/platform/oceandata/
45
the unique persistent identifier (DOI). Data products as concentration and composition of litter 1041 items, of cigarette and fishing related items and plastic bags along the European coasts, density 1042 and composition of litter in the seafloor are made available through the viewing service and 1043 through the product catalogue service. 1044
1045
Figure 16: EMODnet open access marine data, metadata and data products across seven thematic areas. Data and web services offer unique ways to discover, visualize, download and work with marine data.
The Living Atlas of the World 1046
Living Atlas of the World24 is currently the world’s largest GIS digital library that includes a rich 1047 set of thousands of ready-to-use online data layers and maps, as well as related capabilities (e.g., 1048 geocoding, routing, geoenrichment). All assets may be accessed for use in desktop, server, 1049 mobile, and/or web mapping applications. The content is hosted by Esri but the mostly open-1050 access contributions are from scores of partners from government, NGOs, academia, and the 1051 private sector, representing the top 1% of ArcGIS Online’s25 11 million public items, accessed by 1052 1.6 million users daily, and with 4.5 billion map tile requests monthly. The Living Atlas is useful 1053 and reliable for hundreds of topics (e.g., Oceans Chapter of the Atlas)26, including ocean 1054 conservation, coastal and marine spatial planning, ocean resource management and marine litter 1055 surveys. 1056
24 https://livingatlas.arcgis.com/en/
25 http://www.arcgis.com/home
26 https://livingatlas.arcgis.com/en/browse/#d=2&q=oceans&categories=Environment:0110000000
46
1057
Figure 17: The Living Atlas of the World includes over 7000 ready-to-use datasets, maps and apps for empowering many environmental data systems. Partners with and contributors to the Atlas include NOAA, the Marine Conservation Institute, the European Space Agency, NatureServe, GRID-Arendal, and the UN Environment Programme World Conservation Monitoring Centre.
Resource Watch 1058
The Resource Watch27 platform, hosted by the World Resources Institute (WRI), is a free, open 1059 data visualization platform that includes more than 200 data sets on topics ranging from climate 1060 change to agriculture. Data in the platform are curated by WRI experts and extracted from peer 1061 reviewed and verified sources. Resource Watch data visualization functionalities include the 1062 ability to overlay data sets, create dashboards, and download data from the original source. 1063
Earth Challenge 2020 1064
In recognition of the 50th anniversary of Earth Day on April 22, 2020, a consortium of partners led 1065 by the Woodrow Wilson International Center for Scholars, Earth Day Network, and U.S. 1066 Department of State are launching Earth Challenge 2020 as the world’s largest coordinated 1067 citizen science campaign. Earth Challenge 2020 initially focuses on six research areas, including 1068 plastics pollution, and seeks to harmonize existing citizen science data through an open, API-1069 enabled platform and enable new data collection through a mobile application. While the project 1070 will launch in April 2020 with a global outreach campaign, the ultimate goal of this initiative is to 1071 create a long-lasting infrastructure for supporting interoperable citizen science data. Initially, the 1072 project seeks to harmonize and make available a subset of citizen science data on beach/ 1073 shoreline litter collected through NOAA’s Marine Debris Tracker App, EEA’s Marine LitterWatch 1074
27 https://resourcewatch.org/
47
App, Ocean Conservancy’s Clean Swell App, and the Earth Challenge 2020 app. Earth 1075 Challenge 2020 data will be discoverable through GEO DAB, and accessible through GEOSS. 1076
1077
48
Section 5: Proposed Features of a Global Platform and 1078
Required Resources 1079
As noted in Section 3 – marine litter data is diverse and widespread. The lack of standardized 1080 methods and data standards, as well as for the extraction of information and knowledge from the 1081 available data, is currently limiting the understanding of the sources, the transportation 1082 trajectories, the global distribution, the fate, and the impacts of marine litter. Likewise, those 1083 addressing the challenges of mitigating marine litter in all societal sectors have a wide range of 1084 information and knowledge needs. Knowledge needs also change – sometimes rapidly – as new 1085 threats emerge or the extent of environmental impacts is revealed. Thus, linking relevant data 1086 and science to societal decision and policy-making poses a number of challenges to current 1087 approaches in linking data to policies. 1088
In order to build a useful Global Platform for Marine Litter we must address some key questions 1089 related to: 1090
1. Usability — who is it for? 1091 2. Accessibility — who can access it? 1092 3. Capacity — are users equipped? 1093 4. Political buy-in — is there wider support? 1094 5. Governance — frameworks, policies in place? 1095 6. Sustainability — long-term thinking & funding 1096
For the Global Platform to be a useful tool, stakeholders must come together to identify the 1097 priorities in terms of the necessary policy drivers and the corresponding information and 1098 knowledge products to be made available. This white paper provides insights on the high-level 1099 legislative frameworks, the monitoring techniques and observational datasets, the existing 1100 databases, the algorithm and analytics elements, the existing indicators and their technical 1101 readiness level, as well as examples of existing data management and visualisation platforms. 1102 The Global Platform for Marine Litter will need to bring these components together and offer the 1103 relevant insights to its users. 1104
Challenges for integrating and extracting information from data include: 1105
• distributed datasets 1106 • different observation methods, protocols and standards 1107 • disparate temporal and geospatial scales 1108 • metadata standards not set (recommendation for standards) 1109 • quality often unknown 1110 • inconsistent or poorly documented data policies (including open data policies) 1111
Desired Features of the Global Platform 1112
Data Ingestion Portal 1113
49
A relevant part of marine litter data (beach, seafloor, floating micro litter) is regularly monitored at 1114 national, regional or higher level in EU. Nevertheless, a lot of marine litter data is still available 1115 only in peer-reviewed literature or grey literature. In order to encourage publication of data in 1116 openly available databases, a data ingestion portal should be included (see as example EMODnet 1117 Data Ingestion https://www.emodnet-ingestion.eu/). For data upload, ingestion tracking, storage 1118 and transformation this component includes common sets of agreed specifications for data (and 1119 metadata) models, schemas and templates. 1120
Data Brokering Functionality 1121
The use of brokering services (such as the GEOSS platform) to provide data discovery and 1122 access to data for non-core indicator data that is currently distributed across a variety of 1123 databases. 1124
Knowledge resource repository 1125
A central digital archive providing access to codified knowledge, featuring replicable, open 1126 science workflows for using data to support marine litter knowledge extraction in support of 1127 policies. The repository contains documentation linking and providing access to: 1128
• Legislative frameworks, action plans, etc. 1129 • List and roles of stakeholders 1130 • Targets, Indicators, etc. 1131 • Research papers describing methods; 1132 • Monitoring methodologies and algorithms; 1133 • Datasets available (in situ, satellite, airborne, citizen science); and 1134 • Results and scenarios for verification. 1135
Data Processing and Analysis 1136
Computing resources and tools for processing and analysis to extract “knowledge”, such as 1137 change detection, trend analysis, etc. (e.g. change over time in plastic bags as a result of a ban, 1138 information about the sources of marine litter on a particular beach, etc. 1139
Data Visualization 1140
Viewing and publishing services for displaying aggregated (and disaggregated) datasets, maps, 1141 indicators. 1142
Standardization and Interoperability 1143
As described in Section 3, marine litter data is available from different monitoring methods, 1144 protocols and standards, at disparate temporal and geospatial scales, with distinct quality levels, 1145 and with contrasting readiness levels. 1146
It is therefore critical for the Global Platform to adopt consistent methodological standards and 1147 ensure the consistency of data to enable the comparison of indicators at different scales and 1148 between different regions. 1149
50
This includes also the standardisation of marine litter terms and common vocabularies 1150 (semantics). For example, the SeaDataNet pan-European infrastructure for ocean and marine 1151 data management has identified common set of terminologies, metadata attributes, data schemes 1152 and models which are used to uniformly populate the EU EMODnet Chemistry marine litter 1153 database (Addamo et al., 2018; Molina Jack et al., 2019) for the implementation of the Marine 1154 Strategy Framework Directive (MSFD) (European Union, 2008). 1155
The experience of EMODnet Chemistry in the integration of heterogeneous data sources 1156 (collection, standardization, quality control and sharing) started already in 2009, with data related 1157 to eutrophication and contaminants (descriptors 5, 8 and 9 of MSFD). The Chemistry consortium 1158 was already experienced in the data management of physical and chemical oceanographic 1159 information thanks to the activities carried out during the SeaDataNet project. In the last years, 1160 this experience has been expanded with the request to manage marine litter data (Descriptor 10 1161 of MSFD). 1162
Since the beginning, the marine litter data management plan has been to adopt consolidated data 1163 formats when available and adapt them when needed. Following this approach, three specific 1164 methods for beach, seafloor and floating microlitter have been adopted, using the best available 1165 reference documents to develop a tailor-made approach at European scale (Martín Míguez et al., 1166 2019). 1167
The ingestion of the litter datasets has been challenging due to the complexity of the information 1168 and the heterogeneity of the source data. One of the key elements of the success in the data 1169 ingestion was the interaction with data originators. The consortium established a communication 1170 with data originators (direct or through contacts from the Regional Sea Conventions) that allowed 1171 to set up a feedback quality loop done in contact with data originators. This step was crucial to 1172 clarify doubts on reported data and to detect potential duplicates and errors in datasets (or part 1173 of them). 1174
For beach litter, the ingestion of EEA Marine Litter Watch (MLW) datasets is still ongoing. EEA 1175 MLW collects data both from official monitoring and from citizen science. Data from citizen science 1176 represents a really interesting and relevant source of marine litter data due to its wide distribution 1177 (bottom-up approach). However, datasets collected with MLW app, have a strong heterogeneity 1178 in metadata and data quality. Data from monitoring activities or from citizen science can have a 1179 really different quality. For example, for single citizen surveys, it is not mandatory to specify the 1180 identity of the data originator, therefore the feedback quality loop regularly done with known 1181 originator is not possible through traditional methods such as expert review. Additionally, the 1182 identification of surveyed beaches is very relevant for the consolidated monitoring in order to have 1183 time-series data on the same place. Instead, for citizen science data the focus is given to the 1184 survey location, aside from the beach where this survey has been performed. As a result, the 1185 identification of the surveyed beaches along the time can be really difficult. An on-line beach 1186 catalog or an OGC layer providing information (coordinates and metadata) on the surveyed 1187 beaches can help in the integration of official monitoring data and citizen science apps, and 1188 identify areas to prioritize for repeated sampling to promote the collection of time-sensitive data. 1189
The use of citizen science data can be very useful, especially where monitoring programs are 1190 scarce or even not in place. However, the implementation of tools that ensure a minimum quality 1191 of data are needed. One promising area for future work is cross-validating citizen science data 1192 and small aircraft data. A second is the use of machine learning (ML) to identify different types 1193
51
of marine debris photographed by citizen science volunteers as an alternative, or complement, to 1194 expert-based data validation techniques. 1195
Despite all the efforts to harmonize and integrate the available information, the production of data 1196 products able to summarize and highlight specific features is not an easy task. Due to the 1197 complexity and heterogeneity of the surveyed data, their integration is sometimes not possible. 1198 As an example, the use of different gears (nets with different characteristics) for seafloor litter 1199 sampling leads to non-comparable data, due to the differences in the sampling efficiency of the 1200 nets. Any efforts to harmonize and integrate data must also consider data policies, including use 1201 conditions and requirements such as (in the case of citizen science) attribution to a citizen science 1202 volunteer. 1203
Open Platform 1204
To build a customer-centric and user defined Global Data Platform there will be a need to move 1205 from pre-defined data products to content-as-a-service model. Cloud computing and data 1206 analyses between different spatial and non-spatial data is imperative in order to bring together 1207 spatial Earth observation data, socio-economic data to generate knowledge and information (e.g. 1208 indicators) which respond to policy needs. This will require an open, collaborative and federated 1209 platform, where data producers can host, manage and share their own data locally. The platform 1210 must provide open standard interfaces so that information can be exchanged and services can 1211 also be accessed from existing national/regional platforms and systems. 1212
The Global Data Platform will need to tackle a number of challenges to continuously gather, 1213 curate, keep updated, and disseminate actionable information to become really indispensable for 1214 decision making. 1215
- Usability: The platform must be demand-driven and user-oriented. If the objective is to 1216 develop a platform (Knowledge Hub) with highest readiness level: 9 - System, process, 1217 product, service or tool approved for deployment and use in decision making (transition 1218 complete), it is critical to involve key users in the earliest stages of the design. We have 1219 to avoid the risk to implement yet another technology-driven, supply-side data delivery 1220 platform not responding to users need. 1221
- Data vs Knowledge: The platform should be result-oriented, providing clear and objective 1222 guidance to decision makers. Raw data must be transformed into actionable knowledge 1223 to drive policy making and mitigation actions. Assimilative numerical models, intelligent 1224 algorithms, remote sensing and advanced visualization tools may help contextualize the 1225 information and help the development of operational monitoring systems (Atwood et al., 1226 2019). In addition, indicators be must cross-referenced with socio-economic data and 1227 provide scenarios for decision-makers to respond to the challenges of adapting to and 1228 coping with these impacts. Indicators must be contextualized in order to be used 1229 operationally in the decision-making process. For example, is the mean number of MPs/g 1230 ww 3.0 ± 0.9 (SE, n 1⁄4 36) (3.2 ± 0.52 (SE) found per mussel too high? Too small? Not 1231 relevant? Do these numbers have any health impact on wildlife or human beings so the 1232 decision makers can understand and act accordingly? 1233
- A Platform for decision makers: The platform needs to deliver more than quantitative data. 1234 Policy makers primarily adapt or adopt tested and tried policies and “do not target plastic 1235 waste once it has entered the environment; instead they aim to reduce the quantity of 1236
52
plastic production and use, before it is likely to enter the environment. In contrast, waste 1237 abatement outreach programs and infrastructure commonly target plastic waste before 1238 and after it has entered the environment. These strategies try to prevent and remove 1239 plastic waste from entering the environment and prevent coastal deposition” (Willis et al. 1240 2018). It would be very useful for decision makers to have linked to the core monitoring 1241 areas examples of successful policies, regulations, awareness/abatement campaigns and 1242 strategies to prevent and reduce plastic pollution in general, and marine plastic, in 1243 particular. 1244
- One platform for different countries? Decision makers need to have correct insights at the 1245 right scale, however the requirements linked with these insights might vary between 1246 countries. The platform must therefore be configurable and scalable in order for countries 1247 to be able to upload and analyse national data and compare it to broader global context. 1248 As identified by UN Environment (Jensen et al., 2019a, Jensen et al., 2019b), this is 1249 necessary condition if the platform is to generate the correct insights at the right scale, 1250 deliver these at the right time and in the right format in order to influence decision-making. 1251
From Open Data to Open Science 1252
It is widely accepted that the benefits from adopting an open data policy include supporting broad 1253 economic benefits and growth, enhancing social welfare, growing research and innovation 1254 opportunities, facilitating knowledge sharing and effective governance and policy making 1255 (CODATA & Uhlir, 2015). Not all governments have however established national Open Data 1256 regulations/policies to enable agencies to share Earth observation datasets nationally, regionally 1257 and internationally. 1258
Adopting an open data and open access policy is not sufficient on its own, it is important to provide 1259 a platform to maximise the “reproducibility spectrum”, where data, code, analysis procedures, 1260 best practices, and literature, are shared and replicable. In addition, a data policy following the 1261 FAIR (Findable, Accessible, Interoperable and Reusable) principles, will support the development 1262 of solutions which are co-designed between research institutions, societal groups, government 1263 agencies, third sector and industry. These elements will guarantee the implementation of an open 1264 science platform where users are empowered and knowledge is reproducible. Reliable access to 1265 open data is also a necessary requirement for any platform that seeks to support ongoing 1266 monitoring and assessment, including monitoring progress against the SDGs. 1267 Tanhua et al. (2019), outline how these principles apply to ocean data and discuss why ocean 1268 science is an essential foundation for the development of new services made possible with big 1269 data technologies. 1270
Partnerships 1271
Partnerships – public-private partnerships are a critical success factor for the implementation of 1272 the Global Platform. When planning the next steps of this project we must make sure that all 1273 potential stakeholders are included. This – inter-alia – encompasses: end users to identify and 1274 understand (not assume) their requirements, plus observation and monitoring; data management; 1275 big-data analysis and analytics (including AI); computing and geospatial infrastructure, social 1276 science and policy communities. 1277
53
Some technological partners could include: 1278
- The Environmental Systems Research Institute (ESRI; https://www.esri.com/en-1279 us/about/science/initiatives/ocean-science) 1280
- The AI for Earth initiative (https://www.microsoft.com/en-us/ai/ai-for-earth) 1281
- Google Earth 1282
One example of such a public private partnership used make progress on an SDG indicator is the 1283 Water Related Platform28. This is a free platform bringing together the Joint Research 1284 Centre's expertise in satellite data and data analysis, Google's cloud computing and artificial 1285 intelligence and UN Environment’s scientific knowledge. Another example is GEO Blue Planet’s 1286 partnership with ESRI in support of SDG 14.1.1’s eutrophication methodology for summarizing 1287 chlorophyll-a over time in four pilot areas worldwide29 (Smail et al. 2019). Still another is Earth 1288 Challenge 2020, which integrates data from public sector agencies (NOAA and EEA) with NGO 1289 data, and makes integrated data accessible through ESRI’s ArcGIS platform. These examples 1290 demonstrate both the necessity of public private partnerships, and early indicators of collaboration 1291 that could be built on and expanded in a larger, coordinated effort. 1292
1293
1294
28 https://ec.europa.eu/jrc/en/science-update/monitoring-our-blue-planet-first-sdg-indicator-platform-launched-google-jrc-and-un-environment
29 https://storymaps.arcgis.com/stories/37cadf2878e64cd9b34df62baa732b4c
54
Section 6: Marine Litter in a Digital Ecosystem for the 1295
Environment – Thought on A Pilot Project 1296
Section 4 emphasizes the proliferation of data and knowledge platforms that aim at enhanced 1297 data integration and improved access to knowledge derived from data. Most of these platforms 1298 take a thematic approach or are targeting specific user groups. Adding another traditional platform 1299 that aims to serve the users who have knowledge needs related to marine litter would increase 1300 the proliferation. It is very unlikely that adding more of the same will meet the urgent and rapidly 1301 changing needs in the Anthropocene. Considering that the future of humanity and many other 1302 species depends on a well-informed stewardship of the planet, it appears mandatory to make an 1303 effort to fully exploit the wealth of the ever increasing global data resource utilizing leading edge 1304 technologies, approaches and concepts. 1305
A fundamentally different alternative approach has been proposed by Campbell and Jensen 1306 (2019a,b). A “global digital ecosystem for the environment” would utilize the rapid development 1307 of new technologies and methodologies to create an ecosystem of active species interacting with 1308 each other and users (Figure 18). However, a “healthy” ecosystem has a broad diversity of active 1309 species that interact with each other and evolve over time. Developing the concept of an 1310 ecosystem that integrates data, information derived from the data and knowledge co-created in a 1311 collaboration of human agents with the data and information requires to identity the species that 1312 live in this ecosystem. Similar to a biological ecosystem, it is fundamental to recognize the 1313 keystone species that are central to the functioning of the ecosystem and that determine the 1314 nature of this system. 1315
To some extent, the World Wide Web is an ecosystem in which a large diversity of Web-species 1316 interact with each other, compete, benefit from each other, and evolve independently over time. 1317 The digital ecosystem for the environment would have to exhibit similar characteristics in order to 1318 be an ecosystem. The current perception of the world of data, however, does not provide for this. 1319 In general, data are perceived as passive objects that need to be discovered, accessed, and 1320 processed in order to extract information. Progress towards a digital ecosystem would require a 1321 fundamental transition from this current perception of data to a new perception of data as active 1322 subjects (Plag and Jules-Plag, 2019). In this perception data subjects can interact with other data 1323 subjects and human agents in order to actively provide access to the information embedded in 1324 the data. 1325
1326
1327
55
1328
Figure 18: Digital elements that could facilitate a digital ecosystem for the environment. From 1329 Campbell and Jensen (2019a). 1330
In the following, we provide initial thoughts on a pilot project designing and implementing a digital 1331 ecosystem focusing on marine litter. This proposal for the Marine Litter Digital Ecosystem (MLDE) 1332 should be further developed in a participatory approach including the relevant communities. It is 1333 recommended to prepare a white paper that further develops the thoughts provided below. 1334
Initial Thoughts on the Marine Litter Digital Ecosystem 1335
The species in the Marine Litter Digital Ecosystem (MLDE) fall into at least four main domains: 1336
1. Data collection: digital (software) agents that collect new data and generate a flow to 1337 those species that represent data products; 1338
56
2. Data representation: digital agents that represent data objects and can provide 1339 information extracted from these objects, as well as, give access to the data in the objects 1340 and receive feedback on data; 1341
3. Tool representation: digital agents that give access to models and data processing tools; 1342 and 1343
4. Knowledge representatives: digital agents that represent knowledge created by 1344 interaction of human and digital agents. 1345
Thus, the taxonomy of the MLDE would have to include at least the four domains “data collection” 1346 (DCD), “data representation” (DRD), “tool representation” (TRD) and “knowledge representation” 1347 (KRD). It has been suggested to also consider the domain of “best practices” (BPD). Each of 1348 these domains will have classes that consist of a number of families. Each family comprises a 1349 number species. 1350
In the data collection domain (DCD), the classes are defined based on the complexity of the 1351 “sensors” that provide the data stream or data streams. Individual sensors providing a stream of 1352 observations constitute the most basic species. There is considerable diversity ranging from 1353 sensors of physical conditions (e.g., temperature, humidity, displacements, 1354
The agents in the DCD are often linked to sensors or provide means for the reporting of data 1355 through crowd sourcing or the harvesting of data from existing sources. Agents in the DRD that 1356 represent data objects can represent a range of objects from collections of raw data to information 1357 extracted from data, including complex indicators for the environment or synthetic data and 1358 information based on models. If they are asked to provide information that requires data 1359 processing, these agents can interact with the agents in the TRD representing tools for 1360 processing. They would have the full information related to the data object each of them represent 1361 ranging from the actual data, the full metadata, information on quality, usability, former uses 1362 including – to the extent legally allowed – former users, processing tools, and feedback from other 1363 users. 1364
Agents in the KRD representing knowledge have semantic capabilities to answers questions from 1365 human agents. The knowledge they represent is created in collaboration of relevant groups of 1366 societal agents, including scientists, policy makers and other relevant stakeholders. 1367
While most of the agents in the DCD could be reflective agents, the agents in the other domains 1368 would have to be learning agents that combine model-based, goal-based and utility-based agents. 1369 In particular the agents in the DRD and LRD would need semantic capabilities. 1370
A first implementation of the MLDE could utilize the infrastructure available through the Web. The 1371 standardized protocol for the communication between the different agents would require a major 1372 development within the framework provided by the Web. 1373
1374
57
Section 7: Future Developments 1375
In the discussion paper “The Case for a Digital Ecosystem for the Environment”, UN Environment 1376 (Jensen & Campbell, 2019a) makes a compelling case on how data, technology and innovation 1377 can transform the way environmental data are collected and managed, but also how they can 1378 critically enable conditions for better governance. 1379
As reported by the UN Secretary General’s Independent Expert Advisory Group on a Data 1380 Revolution for Sustainable Development,30 without high quality geospatial data, the task of 1381 designing, monitoring, and evaluating effective policies to achieve the Sustainable Development 1382 Goals (SDGs) is almost impossible. For SDG14.1.1, and in particular the marine litter indicator, 1383 new data management technologies, artificial intelligence, cloud computing and cloud storage of 1384 information, together with increased volume of accessible geospatial data, are making it possible 1385 to manage, share, process and analyse large volumes of data in near real time as well 1386 democratizing access to the data itself. 1387
The digital ecosystem proposed by UN Environment would comprise of the following four main 1388 components: (1) data; (2) infrastructure; (3) algorithms and analytics; and (4) insights and 1389 applications. The Global Platform would therefore need to transform data using an underlying 1390 infrastructure combined with algorithms and analytics (see for example artificial intelligence) into 1391 insights and applications that are used by stakeholders (National Statistics Offices, decision 1392 makers, environmental managers, researchers, public & private organisations, citizens, etc.). 1393
As reported by Joppa (2019), to address the challenge of harnessing computing power and 1394 provide actionable solutions for climate change, we need to make use of the three catalysts of 1395 our information age – ubiquity of data, advances in algorithms and access to scalable computing 1396 infrastructure – and apply them to our sustainability challenges. Hence a Global Platform must be 1397 ambitious from a technological perspective and make sure to leverage the data, infrastructure 1398 and algorithms and analytics components, to address the insights and applications as set by the 1399 end-users. 1400
Use of Artificial Intelligence 1401
As mentioned in Section 1, the use of machine learning and deep learning as part of artificial 1402 intelligence (AI) to detect marine litter in the aquatic environment (at the surface and in the water 1403 column) is becoming increasingly relevant. 1404
Fulton et. al. (2014) have evaluated a number of deep learning algorithms performing the task of 1405 visually detecting marine litter, with the objective of exploring, mapping and extracting debris using 1406 autonomous underwater vehicles (AUVs). Balas et. Al. (2004) have applied artificial intelligence 1407 techniques of neural network and fuzzy systems to determine beach litter grading based on litter 1408 surveys. Kylili et. al. (2019) have shown the added value of deep learning techniques in 1409
30A world that counts: Mobilising the Data Revolution for Sustainable Development. http://www.undatarevolution.org/wp-content/uploads/2014/11/A-World-That-Counts.pdf.
58
automatically identifying and determining the amount of floating marine litter, with a success rate 1410 of approximately 86%. Schulz and Matthies (2014) found artificial neural networks eligible to 1411 deliver reliable predictions of marine litter in the southern North Sea with relatively low 1412 computational effort and little input of information. Toro (2019), proposes the use of deep neural 1413 networks to survey and detect marine debris at the bottom of the water column from Forward 1414 Looking Sonar (FLS) images. The automatic detection and quantification of small microplastics 1415 particles (20-1000µm) through fluorence microsocopy and image analysis is helping to address 1416 the difference in marine litter fraction spatial distribution between surface and water column (Erni-1417 Cassola et. al., 2017). 1418
These artificial intelligence based solutions provide fast, scalable, and potentially cost effective 1419 automatic methods for identifying and evaluating marine litter. As more and more remotely sensed 1420 observational data is available from autonomous underwater vehicles (AUV) and remote platform 1421 aerial systems (RPAS) in forms of videos and imagery, machine learning provides a solid, 1422 sustainable and trustworthy alternative to the standard visual-census approach, from both a time 1423 and classification perspective (Martin et. al, 2018 and Moy et. al., 2018). 1424
Integrated Marine Debris Observing System 1425
While there is an existing wealth of data available, observations of marine litter sources, 1426 composition, pathways and distributions in the ocean are sparse and inaccurate. For example, 1427 total amounts of plastic, and other man-made debris in the ocean and on the shore, temporal 1428 trends, degradation processes, vertical fluxes and tie scales are largely unknown (Maximenko & 1429 Corradi, 2019). There is a need to develop a long-term observation platform which is able to 1430 provide the necessary monitoring data for mitigating the impacts of marine litter on the ecosystem. 1431
Two years after the Paris Climate Agreement, the world’s nations mobilized their efforts to tackle 1432 climate change. Space Agencies and other key stakeholders recognized the need to implement 1433 a Space Climate Observatory (SCO31). Along these lines, there is a need to mobilise those 1434 stakeholders committed to making meaningful advances, that the development of an Integrated 1435 Marine Debris Observing System (IMDOS) is critical. Investing in new dedicated space and in-1436 situ programmes, federating existing national and regional databases and coordinating all leading 1437 actors, will offer unified access to a vast majority of marine debris data (acquired from space and 1438 in-situ), which can deliver indicators and decision-support tools (integrating other sources of data) 1439 via a Global Platform for Monitoring Marine Litter and Informing Action. 1440
New Approaches for monitoring harm caused by marine litter 1441
Transoceanic rafting is a fundamental feature of marine evolutionary biogeography and ecology 1442 (Carlton et al., 2017) and the transport of biota on litter items has become a new problem, because 1443 of the recent proliferation of floating particles, which are mostly plastics. Trillions of both micro 1444 and macroplastics at the surface, and sunk debris are all potential carriers of marine organisms, 1445 with advantages for plastic as a transport mechanism in its longevity at sea, its surface properties 1446 favoring attachment and a passive and low speed dispersion. Hundreds of different species, from 1447
31 https://www.spaceclimateobservatory.org/presenting-sco/?lang=en
59
bacteria to larger invertebrates, representing more than 380 taxa, have been shown to settle on 1448 plastics, also on deep sea litter, acting as new habitats. In addition to the alteration the 1449 composition of ecosystems and the possible changes in genetic diversity (Werner et al., 2016; 1450 Carlton et al. 2017), some of them may be at risk like Harmful algal blooms related dinoflagellates, 1451 pathogens to fish and human and invasive species (Werner et al., 2016; GESAMP, 2020). 1452
To date, the settling of species on marine litter, from microorganisms to large invertebrates, 1453 planktonic or benthic, has not been recorded systematically. Because of the risk associated to 1454 their transport, collection of data on rafted organisms, their possible toxicity and mode of invasion 1455 has become critical. In a recent G7 workshop (G7, 2019), it has been concluded that monitoring 1456 should include more knowledge on the microbiology and other species that can present a risk 1457 (invasive, harmful algae, pathogens species) by colonization of plastics, and are subject to 1458 transport on plastics. Sharing data through a dedicated plastics database collecting information 1459 on the colonization of plastics is a priority that will provide historical records, evaluate trends and 1460 support risk assessments. Exploring the various options, including the generation of a new 1461 database or establishing links with existing databases on invasive species, on a global scale 1462 (GISP, http://issg.org/database/welcome/) or at regional level (EASIN, 1463 https://easin.jrc.ec.europa.eu/easin), also considering the possible support of existing database 1464 on marine litter (ICES/PICES, RSCs,EMODNET, etc.) will be part of the future strategy, still to be 1465 defined. 1466 1467
References 1468
Addamo, A., Brosich, A., Montero, M., Giorgetti, A., Hanke, G., Molina, M., & Vinci, M. (2018). 1469 Marine litter database. Lessons learned in compiling the first pan-European beach litter 1470 database. 1471
Andrady, A. L. (2011). Microplastics in the marine environment. Marine Pollution Bulletin, 62(8), 1472 1596-1605. http://www.sciencedirect.com/science/article/pii/S0025326X11003055 1473
ASEAN. (2007). ASEAN Tourism Standards Retrieved from Bangkok, Thailand 1474
Aydin, O. and Butler, K. A., 2019. Predicting global seagrass habitats in Wright, D. J. and 1475 Harder, C. (eds.), GIS for Science: Applying Mapping and Spatial Analytics, Redlands, 1476 CA, Esri Press, pp. 58-69, doi:10.17128/9781589485303 1477
Browne, M. A., Crump, P., Niven, S. J., Teuten, E., Tonkin, A., Galloway, T., & Thompson, R. 1478 (2011). Accumulation of Microplastic on Shorelines Woldwide: Sources and Sinks. 1479 Environmental Science & Technology, 45(21), 9175-9179. Article. <Go to 1480 ISI>://WOS:000296212700011 1481
Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, L. Thomas. 1482 (2001). Introduction to distance sampling: Estimating abundance of biological 1483 populations. 448. 1484
Campbell, J. and Jensen, D. E., 2019. The promise and peril of a digital ecosystem for the 1485 planet, United Nations Environment Programme (UNEP). Available at 1486
60
https://medium.com/@davidedjensen_99356/building-a-digital-ecosystem-for-the-planet-1487 557c41225dc2. 1488
Campbell, J. and Jensen, D. E., 2019. Could a Digital Ecosystem for the Environment Have the 1489 Potential to Save the Planet? National Council for Science and the Environment. 1490 Available at https://www.ncseglobal.org/ncse-essays/could-digital-ecosystem-1491 environment-have-potential-save-planet 1492
1493 Carlton, J., Chapman, W., Geller, J., Miller, J. , Carlton, J., McCuller, M. (2017). Tsunami-Driven 1494
Rafting: Transoceanic Species Dispersal and Implications for Marine Biogeography. 1495 Science, 357(6358), 402–1406. 1496
Choy, C. A., Robison, B. H., Gagne, T. O., Erwin, B., Firl, E., Halden, R. U., et al. (2019). The 1497 vertical distribution and biological transport of marine microplastics across the epipelagic 1498 and mesopelagic water column. Scientific Reports, 9, 9. Article. <Go to 1499 ISI>://WOS:000470243400001 1500
CODATA, & Uhlir, P. (2015). The Value of Open Data Sharing: A CODATA Report for the 1501 Group on Earth Observations. Retrieved from 1502
D’Asaro, E. A., Shcherbina, A. Y., Klymak J. M., Molemaker, J., Novelli, G., Guigand, C. M., 1503 Haza, A. C., Haus, B. K., Ryan, E. H., Jacobs, G. A., Huntley, H. S., Laxague, N. J. M., 1504 Chen, S., Judt, F., McWilliams, J. C., Barkan, R., Kirwan, A. D., Poje, A. C., Özgökmen, 1505 T. M., 2018. Ocean convergence and the dispersion of flotsam. Proceedings of the 1506 National Academy of Sciences Feb 2018, 115 (6) 1162-1507 1167; DOI: 10.1073/pnas.1718453115 1508
Dumichen, E., Eisentraut, P., Bannick, C. G., Barthel, A. K., Senz, R., & Braun, U. (2017). Fast 1509 identification of microplastics in complex environmental samples by a thermal 1510 degradation method. Chemosphere, 174, 572-584. Article. <Go to 1511 ISI>://WOS:000397076000063 1512
European Commission. (2018). Reducing Marine Litter: action on single use plastics and fishing 1513 gear (Commission staff working document, impact assessment). (SWD(2018) 254 final). 1514
European Commission JRC. (2013). Guidance on Monitoring of Marine Litter in Europeans 1515 Seas. Retrieved from https://ec.europa.eu/jrc/sites/jrcsh/files/lb-na-26113-en-n.pdf 1516
European Union. (2008). Directive 2008/56/EC of the European Parliment and of the Council 1517 Establishing a Framework for Community Action in the Field of Marine Environmental 1518 Policy (Marine Strategy Framework Directive) 1519
G7 (2019). Implementing monitoring of plastic and microplastic pollution. SCIENTIFIC 1520 WORKSHOP – G7, Paris, October 14th & 15th 2019, technical annex to the G7 1521 ministerial ( Research ) declaration,2 pages, In press. 1522
Galgani, F., Hanke, G., Werner, S., & De Vrees, L. (2013). Marine litter within the European 1523 Marine Strategy Framework Directive. Ices Journal of Marine Science, 70(6), 1055-1064. 1524 Review. <Go to ISI>://WOS:000323635600001 1525
61
GESAMP. (2019). GESAMP 2019 Guidelines for the monitoring & assessment of plastic litter in 1526 the ocean Reports & Studies 99 (editors Kershaw, P.J., Turra, A. and Galgani, F.) (Rep. 1527 Stud. GESAMP No. 99). Retrieved from http://www.gesamp.org/publications/guidelines-1528 for-the-monitoring-and-assessment-of-plastic-litter-in-the-ocean 1529
Hardesty, B. D., Harari, J., Isobe, A., Lebreton, L., Maximenko, N., Potemra, J., et al. (2017). 1530 Using Numerical Model Simulations to Improve the Understanding of Micro-plastic 1531 Distribution and Pathways in the Marine Environment. Frontiers in Marine Science, 4. 1532 <Go to ISI>://WOS:000457690600030 1533
Jambeck, J. R., Geyer, R., Wilcox, C., Siegler, T. R., Perryman, M., Andrady, A., et al. (2015). 1534 Plastic waste inputs from land into the ocean. Science, 347(6223), 768-771. Article. <Go 1535 to ISI>://WOS:000349221300047 1536
JRC (2012). Life cycle indicators for resources, products and waste: framework. JRC Technical 1537 Reports, Report EUR 25466 EN. European Commission. 1538 https://publications.jrc.ec.europa.eu/repository/bitstream/JRC73336/lbna25466enn.pdf. 1539
Katija, K., Choy, C. A., Sherlock, R. E., Sherman, A. D., and Robison, B.H. (2017). From the 1540 surface to the seafloor: How giant lavaceans transport microplastics into the deep sea. 1541 ScienceAdvances, 3(8), doi: 10.1126/sciadv.1700715. 1542
Kuhn, S., van Oyen, A., Booth, A. M., Meijboom, A., & van Franeker, J. A. (2018). Marine 1543 microplastic: Preparation of relevant test materials for laboratory assessment of 1544 ecosystem impacts. Chemosphere, 213, 103-113. Article. <Go to 1545 ISI>://WOS:000448493700011 1546
Leslie, H. A., van der Meulen, M. D., Kleissen, F. M., & Vethaak, A. D. (2011). Microplastic Litter 1547 in the Dutch Marine Environment. Providing facts and analysis for Dutch policymakers 1548 concerned with marine microplastic litter (report 1203772-000). Retrieved from 1549
Loulad, S., Houssa, R., Rhinane, H., Boumaaz, A., Benazzouz A., 2017. Spatial distribution of 1550 marine debris on the seafloor of Moroccan waters. Mar Pollut Bull. 2017 Nov 1551 15;124(1):303-313. doi: 10.1016/j.marpolbul.2017.07.022. Epub 2017 Jul 24. 1552
Maximenko, N., Corradi, P., Law, K.L., Sebille, E.V., Garaba, S.P., Lampitt, R.S., Galgani, F., 1553 Martinez-Vicente, V., Goddijn-Murphy, L., Veiga, J.M., Thompson, R.C., Maes, C., 1554 Moller, D., Löscher, C.R., Addamo, A.M., Lamson, M., Centurioni, L.R., Posth, N., 1555 Lumpkin, R., Vinci, M., Martins, A.M., Pieper, C.D., Isobe, A., Hanke, G., Edwards, M., 1556 Chubarenko, I.P., Rodriguez, E., Aliani, S., Arias, M., Asner, G.P., Brosich, A., Carlton, 1557 J.T., Chao, Y., Cook, A.-M., Cundy, A., Galloway, T.S., Giorgetti, A., Goni, G.J., 1558 Guichoux, Y., Hardesty, B.D., Holdsworth, N., Lebreton, L., Leslie, H.A., Macadam-1559 Somer, I., Mace, T., Manuel, M., Marsh, R., Martinez, E., Mayor, D., Le Moigne, M., 1560 Jack, M.E.M., Mowlem, M.C., Obbard, R.W., Pabortsava, K., Robberson, B., Rotaru, A.-1561 E., Spedicato, M.T., Thiel, M., Turra, A., Wilcox, C., 2019. Towards the integrated 1562 marine debris observing system. Front. Mar. Sci. 6. 1563 https://doi.org/10.3389/fmars.2019.00447 1564
Martínez-Vicente, V.; Clark, J.R.; Corradi, P.; Aliani, S.; Arias, M.; Bochow, M.; Bonnery, G.; 1565 Cole, M.; Cózar, A.; Donnelly, R.; Echevarría, F.; Galgani, F.; Garaba, S.P.; Goddijn-Murphy, L.; 1566
62
Lebreton, L.; Leslie, H.A.; Lindeque, P.K.; Maximenko, N.; Martin-Lauzer, F.-R.; Moller, D.; 1567 Murphy, P.; Palombi, L.; Raimondi, V.; Reisser, J.; Romero, L.; Simis, S.G.; Sterckx, S.; 1568 Thompson, R.C.; Topouzelis, K.N.; van Sebille, E.; Veiga, J.M.; Vethaak, A.D. Measuring 1569 Marine Plastic Debris from Space: Initial Assessment of Observation Requirements. Remote 1570 Sens. (2019).11, 2443. https://doi.org/10.3390/rs11202443 1571
Martín Míguez, B., Novellino, A., Vinci, M., Claus, S., Calewaert, J.-B., Vallius, H., et al. (2019). 1572 The European Marine Observation and Data Network (EMODnet): Visions and Roles of 1573 the Gateway to Marine Data in Europe. 6(313). Review. 1574 https://www.frontiersin.org/article/10.3389/fmars.2019.00313 1575
Molina Jack, M. E., Chaves Montero, M. d. M., Galgani, F., Giorgetti, A., Vinci, M., Le Moigne, 1576 M., & Brosich, A. (2019). EMODnet marine litter data management at pan-European 1577 scale. Ocean & Coastal Management, 181, 104930. 1578 http://www.sciencedirect.com/science/article/pii/S0964569119304107 1579
NOWPAP CEARAC. (2007). Guidelines for Monitoring Marine Litter on the Beaches and 1580 Shorelines of the Northwest Pacific Region. Retrieved from http://www.cearac-1581 project.org/RAP_MALI/monitoring%20guidelines.pdf 1582
Opfer, S., Arthur, C., & Lippiatt, S. (2012). NOAA Marine Debris Shoreline Survey Field Guide. 1583 Retrieved from 1584 https://marinedebris.noaa.gov/sites/default/files/ShorelineFieldGuide2012.pdf 1585
OSPAR. (2017). OSPAR Intermediate Assessment 2017 Retrieved from 1586 https://oap.ospar.org/en/ospar-assessments/intermediate-assessment-2017/pressures-1587 human-activities/marine-litter/ 1588
Pham CK, Gomes-Pereira JN, Isidro EJ, Santos RS, Morato T (2013) Abundance of litter on 1589 Condor seamount (Azores, Portugal, Northeast Atlantic). Deep-Sea Res. Part II-Top. 1590 Stud. Oceanogr 98: 204–208. 1591
Pham CK, Ramirez-Llodra E, Alt CHS, Amaro T, Bergmann M, Canals M, et al. (2014) Marine 1592 Litter Distribution and Density in European Seas, from the Shelves to Deep Basins. 1593 PLoS ONE 9(4): e95839. https://doi.org/10.1371/journal.pone.0095839 1594
Pinto, C. A., Garvey, P. R., 2013. Advanced Risk Analysis in Engineering Enterprise Systems. 1595 CRC Press, Taylor&Francis Group. 1596
Plag, H.-P., Jules-Plag, S-A., 2019. A Transformative Concept: From Data Being Passive 1597 Objects to Data Being Active Subjects. Data, 4(4), DOI: 10.3390/data4040135. 1598
Prabhakaran, S. (2013). Marine waste management indicators in a tourism environment. 1599 Worldwide Hospitality and Tourism Themes, 5(4), 365-376. 1600 https://doi.org/10.1108/WHATT-03-2013-0013 1601
Richardson, K., D.Haynes, A.Talouli, M.Donoghue. (2017). Marine pollution originating from 1602 purse seine and longline fishing vessel operations in the Western and Central Pacific 1603 Ocean, 2003-2015. Ambio. 2017;46(2):190–200. doi:10.1007/s13280-016-0811-8 1604
63
Rivenson, Y., Wu, Y., Ozcan Y. (2019). Deep learning in holography and coherent imaging. 1605 Light: Science & Applications (2019). https://doi.org/10.1038/s41377-019-0196-0 1606
Ryan, P. G., Moore, C. J., van Franeker, J. A., & Moloney, C. L. (2009). Monitoring the 1607 abundance of plastic debris in the marine environment. Philosophical Transactions of the 1608 Royal Society B-Biological Sciences, 364(1526), 1999-2012. Review. <Go to 1609 ISI>://WOS:000267281600005 1610
Schuyler, Q., Willis, K., Lawson, T. J., Mann, V., Wilcox, C., & Hardesty, B. D. (2018). 1611 Handbook of Survey Methodology Plastics Leakage (developed for CSIRO Global 1612 Plastic Pollution Project) 1613
Smail, E., DiGiacomo, P., Takaki, D., Lance, V., Ramachandran, S., VanGraafeiland, K., Wright, 1614 D., Pisut, D., and Butler, K., 2019. Chlorophyll-a, SDG 14.1, UN Sustainable 1615 Development Goals Breakout, Session 103, Citation 35, Proceedings of OceanObs' 19, 1616 Honolulu, HI, 16-20 September 2019, http://www.oceanobs19.net/sessions/. 1617
Spengler, A. a. M. C. (2008). Methods applied in studies of benthic marine debris. Marine 1618 Pollution Bulletin, 56(2), 226-230. Retrieved from 1619
Stuparu, D., Van Der Meulen, M., Kleissen, F., Vethaak, D., & El Serafy, G. (2015). 1620 DEVELOPING A TRANSPORT MODEL FOR PLASTIC DISTRIBUTION IN THE 1621 NORTH SEA. Madrid: Iahr-Int Assoc Hydro-Environment Engineering Research. 1622
Tanguay, G., Rajaonson, J., & Therrien, M.-C. (2012). Sustainable Tourism Indicators: Selection 1623 Criteria for Policy Implementation and Scientific Recognition. Journal of Sustainable 1624 Tourism, 21. 1625
Tanhua, T., Pouliquen, S., Hausman, J., O’Brien, K., Bricher, P., de Bruin, T., et al. (2019). 1626 Ocean FAIR Data Services. 6(440). Review. 1627 https://www.frontiersin.org/article/10.3389/fmars.2019.00440 1628
Thompson, R. C., Moore, C. J., vom Saal, F. S., & Swan, S. H. (2009). Plastics, the 1629 environment and human health: current consensus and future trends. Philosophical 1630 Transactions of the Royal Society B-Biological Sciences, 364(1526), 2153-2166. 1631 Review. <Go to ISI>://WOS:000267281600015 1632
Topouzelis, K., Papakonstantinou, A., Garaba, S.P., 2019. Int J Appl Earth Obs Geoinformation 1633 Detection of floating plastics from satellite and unmanned aerial systems (Plastic Litter 1634 Project 2018). Int J Appl Earth Obs Geoinf. 79, 175–183. 1635 https://doi.org/10.1016/j.jag.2019.03.011 1636
Turpin V., Remy E., Le Traon, P.Y., 2016. How essential are Argo observations to constrain a 1637 global ocean data assimilation system?, Ocean Sci., 12, 257–274, 1638 https://doi.org/10.5194/os-12-257-2016, 2016. 1639
van der Mheen, M., Pattiaratchi, C., & van Sebille, E. (2019). Role of Indian Ocean Dynamics on 1640 Accumulation of Buoyant Debris. Journal of Geophysical Research-Oceans, 124(4), 1641 2571-2590. Article. <Go to ISI>://WOS:000467950800017 1642
64
Werner, S., Budziak, A., van Franeker, J., Galgani, F., Hanke, G., Maes, T., Matiddi, M., 1643 Nilsson, P., Oosterbaan, L., Priestland, E., Thompson, R., Veiga, J. and Vlachogianni, 1644 T. 2016. Harm caused by Marine Litter. MSFD GES TG Marine Litter Thematic Report. 1645 JRC Technical report. EUR 28317 EN; doi: 10.2788/19937 1646
Woodall, L. C., Sanchez-Vidal, A., Canals, M., Paterson, G. L. J., Coppock, R., Sleight, V., et al. 1647 (2014). The deep sea is a major sink for microplastic debris. Royal Society Open 1648 Science, 1(4), 8. Article. <Go to ISI>://WOS:000209732100006 1649
WTO. (2003). Sustainable Development of Ecotourism: a Compilation of Good Practices in 1650 SMEs. Retrieved from Madrid, Spain 1651
Yu, J. P., Wang, P. Y., Ni, F. L., Cizdziel, J., Wu, D. X., Zhao, Q. L., & Zhou, Y. (2019). 1652 Characterization of microplastics in environment by thermal gravimetric analysis coupled 1653 with Fourier transform infrared spectroscopy. Marine Pollution Bulletin, 145, 153-160. 1654 <Go to ISI>://WOS:000482493000019 1655
Zhu, A., Chen, B., Wang, C., Guo, D., Lei, E., & Jin, F. (2019). Plastic Solid Waste identification 1656 system based on Near Infrared Spectroscopy in combination with support vector 1657 machine. Advanced Industrial and Engineering Polymer Research, 2. 1658
Zulkifley, M. A., Mustafa, M. M., Hussain, A., Mustapha, A., & Ramli, S. (2014). Robust 1659 Identification of Polyethylene Terephthalate (PET) Plastics through Bayesian Decision. 1660 Plos One, 9(12). <Go to ISI>://WOS:000346907600066 1661
1662
