Air Pollution and Sea Pollution Seen from Space | SpringerLink

3.1

Marine Accident and Safety

Of increasing concern in parallel to air pollution are the marine ecosystem degradation and the sea pollution. More than 3 billion people depend crucially on the oceans for their livelihoods. Associated with their presence, it has been reported that litter contaminates marine habitats from the poles to the equator and from the coastal borders to the deep environments. Sea pollution is far from being uniform and that depends on our knowledge of the oceans themselves. Monitoring and forecasting of accidental marine pollution, for which oil spills are an important factor, depend on access to quality information on ocean circulation. It is one of the most important applications of operational oceanography.

Over the past decades, the accuracy of weather and oceanographic forecasts has improved, but millions of dollars of goods and thousands of lives are still lost at sea each year due to extreme weather and oceanographic conditions. In the marine environment, vessels of all sizes are exposed and vulnerable to the elements. High winds, high waves, fog, thunderstorms, sea ice and freezing spray make shipping a high-risk business. Nevertheless, the ocean and seas are a sustainable transport route for the global economy—a “blue economy” estimated at 84% of world trade (The International Union of Marine Insurance Stats report 2018: https://iumi.com). The value of the global ocean economy is estimated at 3–6 trillion USD per year (United Nations Conference on Trade and Development, 2019: https://unctad.org). In addition, ferries carry more than a quarter of the world’s population each year (InterFerry, 2019, see https://interferry.com/ferry-industry-facts/). Maritime incidents endanger lives and property on board and can also cause environmental disasters.

The International Convention for the Safety of Life at Sea (SOLAS), which was created after the Titanic tragedy in 1912 and maintained by the International Maritime Organization (IMO), sets out various standards for the safety, security and operation of ships. IMO and the World Meteorological Organization (WMO), which are specialized agencies of the United Nations, ensure that consistent weather-related marine safety information is available globally for ships at sea. WMO provides a coordination structure that establishes protocols to ensure that coastal countries provide standardized weather information, forecasts and warnings to ensure the safety of people and goods at sea. This includes the requirement to issue explicit warnings for tropical cyclones and for gale and storm wind speeds. Other services include warnings for other severe conditions such as abnormal or devastating waves, reduced visibility due to fog and ice accumulation. Services are prepared and disseminated by National Meteorological Services according to areas of responsibility called Meteorological Areas (METAREA) as shown in Fig. 10. METAREAs are closely aligned with navigation zones, which are used to provide navigational information and warnings to ships at sea. Weather-related maritime safety information is transmitted to ships at sea by countries through the WMO Maritime Broadcast System, which supports the IMO Global Maritime Distress and Safety System (GMDSS).

Fig. 10

Global map of the METAREAs with issuing services

Full size image

These warning and forecasting services are widely used by all mariners, safety and security organizations and economic sectors that make informed decisions using marine weather information to improve the safety of people and goods, environmental safety and socio-economic benefits in marine and coastal environments. Numerical weather, wave and ocean prediction models provide marine forecasters with indications that underpin the forecasting process. Weather and ocean forecasting models are run two to four times a day using the data collected to make future forecasts. These models are based on supercomputers that perform calculations based on complex systems of differential equations that approximate the physical processes of the atmosphere, the oceans and sea ice. Experienced marine forecasters evaluate all available observations and numerical forecasts to ensure that the atmosphere and the ocean are represented as accurately as possible in the initial state of the model and to determine where the greatest model uncertainties lie on a given day. Forecasters then prepare and disseminate their marine weather forecasts, warn of hazardous weather conditions as required and send their forecasts for broadcast on the Internet, radio or satellite.

As a result of decades of coordination with universities, research centers, research infrastructures and private companies, operational oceanographic services, such as the Copernicus Marine Environment Monitoring Service (CMEMS: http://marine.copernicus.eu), have been set up. The combination of advanced observations and forecasts offers new downstream service opportunities to meet the needs of heavily populated coastal areas and climate change. Over the next decade, the challenge will be to maintain ocean observations, monitor small-scale variability, resolve sub-basin/seasonal and interannual variability in circulation, understand and correct model biases, and improve model data integration and overall forecasting for uncertainty estimation. Better knowledge and understanding of the level of variability will enable subsequent assessment of the impacts and mitigation of human activities and climate change on biodiversity and ecosystems, which will support environmental assessments and decisions. Other challenges include the extension of value-added scientific products to socially relevant downstream services and engagement with communities to develop initiatives that will contribute, inter alia, to the United Nations Decade of Ocean Sciences for Sustainable Development, thereby helping to bridge the gap between science and policy.

Forecasting services can play an important role in both incident decision making and the design of response services. Monitoring, forecasting and, to some extent, detection of marine pollution depend mainly on reliable and timely access to environmental data, observations and predictions products. These products provide an overview of the current and future state of meteorological and oceanographic conditions. They can also be used to drive pollutant fate prediction models, either directly or by providing boundary conditions for high-resolution nested models. Access to large geophysical datasets that are interoperable with regional and sub-regional observation and modeling systems, using standard service formats and specifications such as those provided by CMEMS, is an undeniable achievement.

Oil spill prediction is generally performed using a numerical model to predict the advection and weathering of oil in the sea (examples will be discussed in the next section). Although the treatment of these processes can vary considerably from one model to another, they all critically depend on geophysical forcing to determine the fate of the oil spill, particularly its movement implying currents and winds. Oil spill models consider different forcing fields from data on wind, ocean currents, wave energy, Stokes drift, air temperature, water temperature and salinity, turbulent kinetic energy, depending on the settings used by the particular model. For the prediction of the movement of oil slicks on the high seas, ocean circulation data has the greatest potential for improvement, mainly because ocean forecasts are less mature than weather and wave forecasts. Operational ocean forecasting systems offer the promise of improved forecast accuracy through the assimilation of available ocean observations. The geographical scope of these systems extends from basin to global scale, facilitating global oil spill modeling capabilities. It should be noted at this point that oil spill modeling systems can use ocean forecast data in two ways: as direct forcing to the oil drift model and as boundary conditions to higher-resolution local ocean models which, in turn, provide forcing data to the oil drift model.

The forecasting systems used to assist search-and-rescue operations are very similar to those used for oil spills. Speed of response is an essential criterion for finding people lost at sea. The size of the search area grows rapidly over time, and it is important to have accurate environmental data (winds and currents) to reduce this area and quickly find the target you are being looked for. Futch and Allen (2019) state that 60% of SAR incidents under the US Coast Guard areas of responsibility are outside areas where high-resolution wind and current data are available, and only global forecasts are available. Finding solutions that reduce the size of these areas of intervention is therefore crucial.

3.1.1

The Example of the Grande America Accident

The Grande America accident is a recent example of the use of ocean current data from several models. On March 10, 2019, a fire broke out on board the Italian merchant ship Grande America. The ship carries 365 containers, 45 of which contain dangerous goods and 2000 vehicles (cars, trucks, trailers, construction machinery) in its car decks. It sank on March 12 at a depth of 4600 m, 350 km off the French coast, with about 2200 tonnes of bunker fuel on board. More than 1000 tonnes of heavy fuel oil were released into the marine environment on the day the ship sank, followed by 35 days of continuous leaks before the breaches in the hull were filled by an underwater robot. A technical committee bringing together experts from Cedre, Météo-France, IFREMER and SHOM assessed the drift forecasts provided by the MOTHY oil spill drift model (Daniel 1996). The Drift Committee was responsible for providing the maritime authorities with consistent and relevant information on oil drift, observations and forecasts on a daily basis. MOTHY was used daily during the aerial surveillance and recovery period using oceanic forcing fields provided by CMEMS. Significant discrepancies were found between ocean models. Finally, the drift predictions closest to the observations were obtained with the currents provided by the ocean model of lower resolution (Fig. 11). This shows the importance of having several ocean models and the difficulty of using high-resolution ocean forecasts that generate many eddies whose precise location is sometimes difficult to model.

Fig. 11

MOTHY output from a simulated 10-day leak from the Grande America wreck (dark green star near 46° N–6° W). Black disks are aerial observations. Particle drift: in green without ocean current data, in red using ocean current data from the operational Iberian Biscay Irish (IBI) Ocean Analysis and Forecasting system at 1/36 degree, in blue using ocean current data from the operational Mercator global ocean analysis and forecast system at 1/12 degree

Full size image

This example highlights the main challenge for data providers in the future, namely improving the accuracy of current forecasts. Therefore, international collaboration should continue to consolidate work on validation measurements and model comparisons to ensure that a minimum set of measurements is implemented internationally. In particular, it is necessary to include spatially explicit estimates of uncertainty, both for forcing data and for the results of the oil spill model. From this point of view, ensemble forecasting is a very promising research area for quantifying the uncertainties inherent in drift prediction at sea. It should be further studied. Products should be delivered to users efficiently and should be provided with adequate spatial and temporal resolution. The interaction between users and production must be an important criterion for future developments.

3.2

Oil Spill Pollution

The end of the last century and the beginning of this new one have been plagued by major environmental catastrophes in terms of ocean oil pollution. Amoco Cadiz (1978), Exxon Valdez (1989), Sea Empress (1996), Erika (1999), Prestige (2002) and Tasman Spirit (2003) are but examples of oil tanker accidents. Nevertheless, these tanker accidents account for only a maximum of 2–3% of the whole oil pollutions worldwide. Another small percentage (2–3%) is due to platform accidents (Deep Water Horizon is a good example) leaving the rest to wild discharges (see Fig. 12 extracted from Pavlakis et al. 2001).

Fig. 12

Figure from Pavlakis et al. (2001)

Fingerprints of illicit vessel discharges detected on ERS-1 and ERS-2 SAR images, during 1999 within the Mediterranean Sea.

Full size image

Detecting and tracking oil spills were necessary, and rapidly, the best and most efficient means of observation and analysis were recognized through synthetic-aperture radar (SAR or imaging radar) measurements. Indeed, on board satellite or aircraft, SAR is an adapted tool for monitoring the surface of the ocean and therefore for detecting sea surface pollution. Due to microwave properties, SAR measurements do not depend on weather or cloud cover, and slicks have significant backscattering behaviors that can be highlighted by SAR systems. Imaging radars measure mainly the surface roughness and because slicks are producing a strong impact on short waves they modify seawater viscosity and dampened backscatter returns. Nevertheless, automatic spill recognition is not an easy task as the presence of extended dark manifestations in the image, due to occurrences other than spills but comparable to them, yields similar SAR signatures.

These extended dark patches are due to other phenomena also modifying the sea surface conditions: wind, sea state, sea temperature, rain showers, and currents (Girard-Ardhuin et al. 2003). An experienced image interpreter is essential to sort out irrelevant disturbances linked to natural phenomena. Furthermore, specific SAR parameters, using different SAR modes, are to be considered and provide more flexibility for the analysis: radar wave polarization, wavelength, and incidence angle for example, but also several other parameters such as satellite or aircraft flight direction, waves and wind directions. Then, many of the possible confusions arising from automatic recognition can be smoothed by integrating this information surplus, helping to discriminate between different sources of sea surface disturbances.

In order to go beyond oil spill detection and tracking, and for setting an efficient prevention system, one must also detect and recognize the ships responsible for the oil discharges. For that purpose, synergetic approaches have been used, mixing radar imaging inputs with met information and mandatory ship identification systems (AIS—automatic identification system) mainly. This effort is quite important as the vast majority of these pollutions occur near the coastal zones where 80% of the world population lives.

3.2.1

Effect of Slicks on the Ocean Surface and Radar Imaging Measurements

Slicks have the effect of damping the waves at the surface (Alpers and Hühnerfuss 1988, 1989; Solberg et al. 1999). When a slick covers the surface, wind has a lesser effect, and the amplitude of wave crest/trough decreases, implying a surface stress gradient, with opposite strength to this alternated motion. Therefore, a film at the ocean surface implies an energy decrease with distribution to high and short waves, corresponding to the important surface stress decrease. Because the radar is only sensitive to the surface roughness, several other modifying aspects need to be considered: waves, sea properties (diffusion, mixing, turbulence, air–sea exchanges, biology, etc.), surface currents, eddies, internal waves, oceanography, meteorology, atmospheric dynamics, etc., in order to understand the all influences on slick dissipation.

The SAR technology is one of the most suitable (see for instance Girard-Ardhuin et al. 2003, among many others) instruments for oil slick measurements since it does not depend on weather conditions (clouds) nor sunshine, which allows showing illegal discharges that most frequently appear during night; it can also survey storms areas, where accident risks are increased. Added to local aircraft tracking, SAR is helpful for synoptic oil spill monitoring. Its spatial coverage can be as wide as 500 × 500 km, with high-resolution images of about 10 m, or in very high-resolution mode down to a meter-scale resolution with a much smaller coverage (20 × 20 km).

Figure 13a shows an example of a large extent zone North West of Spain, on the path of the Finisterre TSS (traffic separation scheme) used by the oil tankers going back and forth between Rotterdam and Athens. The radar image is acquired by Radarsat-2 on 03/25/2016 at 06:41 UTC. This image was received, processed and analyzed by the VIGISAT 24/7 operational center in Brest, dedicated to the near real-time maritime surveillance services. Figure 13b shows, in the enlargement, not only the slick but several ships as well (white dots).

Fig. 13

a Radar image showing an oil discharge (blue square), b zoom in on the oil slick (Radarsat2@2016)

Full size image

Surface roughness, linked by gravity–capillarity waves, is dampened by slicks, and the radar backscattering level is therefore decreased, yielding dark patches with weak backscattering in comparison with the surrounding regions in radar image, with at least 3 dB contrasts. From a synthesis of experimental previous studies, the best parameters to use to detect slicks are a function of radar configuration, slick nature and meteorological and oceanic conditions. SAR parameters are particularly of importance as their selection may affect the oil slick detection capabilities. Per construction, one can choose the radar wavelength, the radar polarization or the incidence angle of the radar wave. Other parameters strongly affecting the end results are conjunctural, such as the angle between flight direction and wind direction and, of course, the nature of the slick itself. The influence of meteorological and oceanic conditions has been discussed before and is entering into the category of additional parameters allowing some possible discrimination (Girard-Ardhuin et al. 2003).

For two decades (1980–2000), several experiments were conducted with a clear choice for higher microwave frequencies (C- to Ku-bands) over longer wavelength (S- or L-band), and an ability to show about 5 dB contrast for a slick made with “light” fuel, and 10–15 dB contrast for a “heavy” fuel one. Wind speed being an external parameter to consider, C-band frequency seems to be the most suitable frequency allowing to measure strong contrasts up to wind speed about 13 m/s. Wind speed intervenes in the choice of radar polarization. Again C-band frequency seems to be the most suitable choice with VV polarization, notably with strong wind, higher than 11 m/s. Finally, the nature of the slicks plays also a role in detection and, potentially, identification. Indeed, backscatter damping is a function of the origin and of the slick viscosity and elasticity properties. Measured with C-, X- and Ku-bands, wave damping is more important for oil slicks than for natural films, and the latter are often detectable only at low wind speed (< 5 m/s). Thus, using multi-frequency radar measurements is a good solution, when possible, for determining slick nature.

3.2.2

SAR and AIS Coupled Detection

SAR imagery provides a great potential in observing and monitoring the marine environment. Ship detection in SAR imagery is well advanced (Crisp 2004), and knowledge about vessels position and type yields a wide range of applications, such as maritime traffic safety, fisheries control, illegal discharges and border surveillance. Applications developed in this area are vast, and the evolution of SAR sensors capabilities and the availability of large SAR image dataset coupled with cooperative positioning data (such as AIS—automatic identification system ones) allows a much finer monitoring for man-made oil pollution (Pelich et al. 2015a). The AIS system allows the tracks of all vessels (Fig. 14) in the vicinity of the spill (AIS data from 01:00 to 06:47 UTC).

Fig. 14

(courtesy by CLS)

Oil slick highlighted with added information on AIS ship positioning and wind speed and direction. Blue color: AIS positions are 6 h before the SAR image acquisition time. Red color: AIS positions at the SAR image acquisition time

Full size image

In order to identify the potential polluter, a series of steps must be applied. As too many candidates are present, there is a need to reduce the selection of potential ships. Figure 15a then shows the track of only three ships sailing southbound, in the vicinity of the spill. One AIS track is more likely to be the polluter. In order to assess the correct track, a drift algorithm is run for confirmation.

Finally, to assess which track is more likely to point to the polluter, a drift model is applied on all three tracks with an estimation of “fake” pollutions, as if all three ships had polluted. From the AIS-based locations of vessels, and considering the wind and the current information, the displacement of the pollution at sea is computed for a specific time interval (typically 6 h). The pollution transport model is based on a 2D Lagrangian advection scheme which linearly combines wind and current effects, with an additional diffuse scheme (Pelich et al. 2015b; Longépé et al. 2015). Figure 15 shows that only one AIS track is consistent with the oil spill shape.

Fig. 15

(courtesy by CLS)

a Tracks of three main candidates as polluters, b only one track is consistent with the oil spill

Full size image

In summary, ocean oil pollutions are mainly coming from illegal discharges and one of the best tools for detecting them is the imaging radar on board satellites. But SAR instruments cannot, alone, achieve the level of recognition expected by organizations such as European Maritime Safety Agency (EMSA). More evidences are required, and we showed here that by adding informations directly derived from the SAR image (ship detection, wind speed and direction, image interpreters) coupled with external ones (met conditions, AIS positions, drift model), we are able to accumulate enough proofs for having a significant impact on the polluters, including prosecution (a set of evidences can be accepted as a proof).

3.3

The Problem of Plastics as Marine Debris

World plastic production has increased exponentially in recent decades, from 2.3 million of tonnes (Mt) in 1950 to 162 million in 1993 to 448 million by 2015. As a consequence of their high durability (and associated low cost), substantial quantities of end-of-life plastics are accumulating in the environment. Presently, the part of unmanaged debris and plastics that end up in the World Ocean is estimated between 8 and 15 Mt per year. Nevertheless, the real dimension of the problem emerged only recently, and it has been pushed forward the scene under the auspices of the Sustainable Development Goals on life below water (SDG 14) that goals to “conserve and sustainably use the oceans, seas and marine resources for sustainable development.” Before that step, scientific research started in the early 1970s with the report of plastics on the seafloor and at sea and with numerous studies focusing on their impact on a variety of marine animals (Ryan 2015). Marine debris and litter represent nowadays a growing environmental problem and plastic, the largest component of litter, is now widely reported within the marine and as well as all natural environments on Earth. The problem is so huge that it represents threats to the environment, the economy, and human well-being on a global scale (Nielsen et al. 2019). Despite its many benefits on land and extensive usage inside our societies, the plastics and other artificial debris are increasingly being recognized as the source of severe environmental problems, and in particular in the oceans. Among the numerous aspects of the plastic problem, the fact that plastics and debris could travel across large distances, typically at the scale of ocean basins, is at the heart of the research priorities addressed by the scientific community to the G7 Science Ministers at Berlin in October 2015 (Williamson et al. 2016). In the top research priorities listed by this latter report, understanding the sources and the pathways in order to “establish connections between sources and sinks for different types of debris, and how these are influenced by oceanic and atmospheric dynamics” are the two first basic points that need to be solved to define the extent of the problem. More or less, all the different international reports dedicated to the plastic problem at global scales addressed the same starting points in addition to the fact that present ocean circulation models are not able to accurately simulate drift of debris because of its complex hydrodynamics (Maximenko et al. 2019). It is also important to note that, despite some progress made in spectral and hyperspectal remote sensing technologies (Garaba et al. 2018; Goddijn-Murphy and Williamson 2019), observations from space remain limited.

In the following, a simple Lagrangian assessment of dispersion from modeled surface current trajectories at the global ocean scales is addressed. The reason for that is linked to focus more on the physics of the oceanic environment rather than on the material itself requiring to specify, in such a case, different parameters like its density, size, shape and so on. Considering the problem of microplastic marine particles, Khatmullina and Chubarenko (2019) recently conclude that further integration or coupling with different type of models is required to advance on such complex problem. Here, we consider that plastics and marine floating debris represent one unique tracer that provides the opportunity to learn more about the physics and dynamics of the oceans across multiple scales. The evolution of a tracer in any turbulent flows is given by the paradigm of Eckart (1948): At first, during the stirring phase, the variance of the scalar gradient is increased, and later, during the mixing phase the molecular diffusion dominates and the strong gradient disappears (homogeneous final state). However, the “stirring and mixing” problem in a stratified ocean relies on the physics that need to be parametrized in ocean models and great challenging open problems remain at all levels, from very fundamental to highly applied aspects (Müller and Garrett 2002).

For centuries, surface currents were inferred from bottles and drifting objects but, since the 1980s, the World Climate Research Program initiated the global array of surface drifters (with anchors drogue set at 15 m depth). Accumulation of observations and continuous improvement leads to a quite precise time-mean circulation resolving details such as the cross-stream structure of western boundary currents, recirculation cells, and zonally elongated mid-ocean striations (Laurindo et al. 2017). If the averaged seasonal climatology is quite pertinent, daily and near real-time observations are only based on 1200–1300 active buoys on global scales representing approximately ~ 80% of 5° x 5° coverage within the 60 °N–60 °S band. On the other hand, satellite remote sensing proposes surface geostrophic currents derived from the altimetry since the 1990s, and with the addition of a wind-driven Ekman current, one estimate of surface current could be delivered (i.e., Sudre et al. 2013). If the variability of large-scale currents (200-km wavelength and 15-day period) is well depicted by satellite altimetry and vector winds from remote sensing, it leaves important observation gaps that will require other technology such as the understanding of Doppler properties of radar backscatter from the ocean (Ardhuin et al. 2019a, b). Representing another important tool in climate services, ocean reanalyses combine ocean models, atmospheric forcing fluxes, and observations using data assimilation to give a four-dimensional description of the ocean (i.e., Storto and Masina 2016). Based on these estimates, the dispersion dimension of the problem should be addressed from a Lagrangian point of view.

Complementing the traditional Eulerian approach, Lagrangian ocean methodology is a powerful way to analyze the ocean circulation or dynamics. The purpose is based on large sets of virtual (passive or fictive) particles whose displacements (trajectories) are integrated within the 3D/2D and time evolving velocity field. Among the expectations, questions about pathways and flow connectivity can be addressed and it avoids the recourse to the classical use of passive tracers such as geotrace gases. In a simplified framework, the analyses include the evaluation of the ocean currents, the determination of the initial positions of the particle and the computation of trajectories under physical assumptions, and finally their post-treatment. Some concrete examples that have been studied during the recent couple of years include: the tracking of the origins of plastic bottles across the Coral Sea (Maes and Blanke 2015), the discovering of “exit routes” near the core of the subtropical convergence zones of the Pacific Ocean (Maes et al. 2016), the determination of the mesoscale features characterizing the dynamics of the Coral Sea with its impact on water masses and circulation (Rousselet et al. 2016), the potential connection between the South Indian and the South Pacific subtropical convergence zones revealing that large-scale convergence zones should not be seen as “closed systems” (Maes et al. 2018), and more recently, the important role of the Stokes drift in the surface dispersion of floating debris in the South Indian Ocean (Dobler et al. 2019). Based on these different studies and other consistent results reported in the literature, we were able to identity these important findings: We need to know the time variability of surface currents to adequately address the long distance connectivity and pathways, we need to consider high-resolution horizontal currents (mesoscale at least) to estimate the time transfer on specific pathways, and finally, we need to evaluate more accurately the emission scenario (and the sinks) of surface floating litter to access the distributions at global scales.

Regarding the latter aforementioned point, additional experiments were performed using two different scenarios of particle release: The first one is based on one instantaneous release over the full ocean domain of the model (i.e., one particle on each ocean grid point at the initial timestep), whereas the second one is based on a permanent release along the coastlines (i.e., the ocean grid point adjacent to the land mask), the distribution being based on the population density as used by van Sebille et al. (2015). Not surprisingly, both experiments exhibit the presence of the subtropical convergence zones mainly due to the ocean dynamics (Fig. 16), which results in accumulation in the absence of vertical mixing and potential sinks. Nevertheless, the experiment with continuous release along the coastlines shows much more complex structures and higher concentrations in some areas including the different marginal Seas of the western Pacific Ocean, the Indonesian Seas, the Mediterranean Sea and the Bay of Bengal as well as the South Indian Ocean. It is also clear that some places within the tropical band such as in the Eastern Pacific Ocean near the Galapagos Islands or in the Gulf of Guinea are also characterized by some convergence of particles (Fig. 16b). Such places could not be identified in the first experiment (Fig. 16a) due to the divergent nature of the ocean dynamics, whereas they could be seen as semi-permanent places in the second case. To determine which one of these spatial distributions represents the actual concentration of marine debris or plastic at sea is quite impossible, due to the absence of in situ observations, and they are probably not realistic either. It is important to remember that most of the present studies put their focus on the surface layer (and so, on floating material), whereas the vertical distribution of floating plastic depends not only on the particle’s buoyancy, but also on the dynamic pressure due to vertical movements of ocean water. Understanding of the vertical flow in the ocean is challenging because it is induced by several processes acting at different temporal and spatial scales, not always fully resolved by present ocean models. Secondly, most of the present studies have assumed a steady, non-changing ocean circulation. However, low-frequency variations, such as the El Niño Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO), modify the ocean circulation and modulate the different physical processes. Trends associated with climate change can also amplify some of the dispersion processes at work in the future. It is not clear how these processes and the coupling with the internal variability of the ocean will or not affect the dispersion of marine debris and plastics at the global scales.

Fig. 16

Number of particles per one fourth degree cell resulting from the initially instantaneous experiment (a) and from the permanent release along coastlines assuming a lifetime of 20 years for the particles (b) after 29 years of dispersion

Full size image