Macroplastic Storage and Remobilization in Rivers

: The paper presents a conceptual model of the route of macroplastic debris (5 > mm) through a fluvial system, which can support future works on the overlooked processes of macroplastic storage and remobilization in rivers. We divided the macroplastic route into (1) input, (2) transport (3) storage, (4) remobilization and (5) output phases. Phase 1 is mainly controlled by humans, phases 2–4 by fluvial processes, and phase 5 by both types of controls. We hypothesize that natural characteristics of fluvial systems and their modification by dam reservoirs and flood embankments construction are key controls on macroplastic storage and remobilization in rivers. The zone of macroplastic storage can be defined as a river floodplain inundated since the beginning of widespread disposal of plastic waste to the environment in the 1960s and remobilization zone as a part of the storage zone influenced by floodwaters and bank erosion. The amount of macroplastic in both zones can be estimated using data on the abundance of surface- and subsurface-stored macroplastic and the lateral and vertical extent of the zones. Our model creates the framework for estimation of how much plastic has accumulated in rivers and will be present in future riverscapes.

Previous works suggested that both natural (e.g., hydrological processes, channel morphology, and riparian vegetation) and anthropogenic characteristics of river catchments (e.g., plastic waste management) control the input and transport of plastic debris in fluvial systems [12,13,15,[22][23][24][25][26][27][28][29], and that macroplastic constitutes most of riverine plastic debris in terms of mass [26]. Recent studies show that macroplastic debris may be stored on the surface of alluvium and in riparian vegetation (surface storage) and in river sediments (subsurface storage) [21,30]. Moreover, stored macroplastic debris may be remobilized by floodwaters and bank erosion [21,22] and its fragmentation constitutes the main source of secondary microplastic in rivers [12,22]. Taking into account long preservation of macroplastic debris in the natural environment [1,12], its storage-remobilization cycles in fluvial systems may last for decades or centuries [e.g., 19,21]. This implies that the presence of riverine macroplastic and related environmental risk may continue in the future even when an input of new plastic debris to the fluvial systems is decreased. Processes of the storage and remobilization of macroplastic are overlooked in the literature but their understanding is fundamental for: (i) assessment of the amount of plastic accumulated in river systems, (ii) effective targeting the cleanup actions [15], (iii) management of its present and future remobilization, (iv) assessment of related risks, and (v) its influence on the aesthetic value of river landscapes [31][32][33].
In this perspective, the key knowledge to gain in future works is the precise information about the spatial and temporal extent of the storage of macroplastic debris in river deposits (both surface and subsurface) and riparian vegetation and estimation of the potential for its remobilization caused by water inundation and erosion processes. Recent works showed that macroplastic debris can be detected through visual analysis of present and past sediments [1,21,34] and also monitored and mapped with remote sensing methods (e.g., aerial photos, UAV surveys) [35][36][37][38][39][40][41], which creates a promising methodological perspective for future research. However, a conceptual framework and a theoretical scheme for further works on macroplastic storage and remobilization in fluvial systems are not available.
We developed a framework to support future studies on macroplastic storage and remobilization in river systems. In this paper, we present and discuss a conceptual model of macroplastic route through a fluvial system, describing processes of macroplastic input, transport, storage, remobilization and output. We use this model to systematize main controlling factors for each phase and summarize existing methodological approaches which may be used for their quantification in future works. Our goal is to support systematic development of future hypotheses on the processes of macroplastic storage and remobilization in rivers.

Plastic as a new, artificial type of river load
The existing paradigm in hydrological and geomorphological sciences divides sediments transported by rivers (so called river load) into mineral and organic types. However, a growing body of evidence suggests that plastic debris may be treated as a new kind of sediment particles and thus a new type of river sediment load because of its ubiquity, longevity, and abundance in the environment [19]. The processes of transport, deposition and remobilization of mineral and organic sediments are controlled by their physical properties (e.g., density, size) and their relation to river size and hydrodynamics [42][43][44][45]. Recent insights into interactions of macroplastic debris with fluvial processes documented by laboratory experiments [46][47] and monitoring of floating [24,26,27] and deposited macroplastic debris [48] suggest that transport, storage and remobilization of macroplastic debris are also partly related to fluvial processes. The main similarities between macroplastic and larger particles of organic debris (e.g., seeds, vegetation propagules, large wood pieces) encompass their relatively low density and large surface area determining their transport by water in flotation [see 45]. Previous works also indicated a dependence of macroplastic transport on river hydrodynamics [27] and types of aquatic vegetation [26,30]. For example, van Emmerik et al. [26] detected a seasonal trend of changes in macroplastic transport related to the seasonal cycle of water hyacinths growth in the Saigon river, Vietnam. However, in the Seine (France) plastic transport was found to increase ten-fold during high discharge in comparison to low discharge [27]. Bruge et al. [48] found that more artificial debris was deposited on riverbanks in a river section characterized by stagnant water, low channel slope and large-size riparian vegetation, highlighting the importance of flow hydrodynamics and river morphology as potential controls on macroplastic storage. Recently, Cozzolino et al. [30] found that macroplastic abundance on the surface of vegetated areas of a tidal zone was higher in comparison to the adjacent unvegetated areas. This suggests that a similar pattern may develop also along densely vegetated river banks and delta-backwater zones of dam reservoirs.
A difference arises when processes of the input of plastic to a river are compared with those for organic debris [cf. 44,49]. The former seem to depend on both anthropogenic (e.g., urbanization, population density, waste management, road density) [29] and natural factors (e.g., wind, surface runoff), while the latter mainly on natural processes (bank erosion, windstorm, landslides) [cf. 44,49]. However, once macroplastic started to be transported and stored in a fluvial system, natural (e.g., climate, catchment topography, valley and channel morphology) and anthropogenic (dams, embankments) characteristics of the fluvial system seem to influence the transport and storage of macroplastic in a similar way as it was documented for woody debris [45].
We used the above highlighted similarities and information from the literature to infer the main controls on macroplastic storage and remobilization in rivers.

Conceptual model of macroplastic routing through a fluvial system
We divide the route of macroplastic debris through a fluvial system (or its given part, e.g., subcatchment or river reach) into five phases: (1) input, (2) transport, (3) storage, (4) remobilization and (5) output ( Figure 1). Then we assign natural or anthropogenic controls to particular phases of the macroplastic route ( Figure 2). Using this systematization, we demonstrate a change in main controls between different phases of the macroplastic route: phase 1 (input of macroplastic) is mainly controlled by anthropogenic factors, phases 2-4 by natural ones, and phase 5 (output) by both types of factors (Figures 1-2). In Table 1 we summarized methods allowing for measurements of macroplastic abundance in sediments, which may be applied for further quantification of phases 3 and 4.

Input of macroplastic to fluvial system
The input of macroplastic is defined as artificial or natural placing of plastic items within the zone of active fluvial processes (channel, floodplain, eroded terraces), allowing for their immediate (input to flowing water) or delayed (e.g., input to floodplain zone and subsequent transport by floodwater) transport by river (phase 2) ( Figure 1). Input of plastic debris to fluvial systems seems to 5 of 14 be primarily controlled by humans who may dispose litter, including plastic, either scattered in a channel, on river banks or floodplain, or concentrated in dumping sites and landfills located in the riparian zone of a river. At the catchment scale, urbanization, population density and waste management are the most important sources of macroplastic input to rivers [e.g., 29,50,51]. At the river reach scale, the location of dumping sites on river floodplains is an important control on a local input of plastic debris [e.g., 52]. Natural processes such as wind or mass movement processes (e.g., landslide) may also act in a similar way [12], but their importance as input factors seems to be lower than that of humans, especially in populated areas and where dense riparian vegetation occurs along a river course. At the catchment scale, the input of plastic can be indirectly assessed based on natural and anthropogenic characteristics of the area [e.g., 29]. At smaller spatial scales, the location and extent of input zones can be directly mapped during field surveys [e.g., 52] and from remote sensing material such as aerial photos taken during low-altitude drone flights [e.g., 35].

Macroplastic transport
Transport of macroplastic begins when its particles start to be moved by river flow after their input to active river zone or when stored macroplastic is remobilized by floodwater or bank erosion. The occurrence and intensity of this phase depends primarily on fluvial processes (e.g., flood occurrence, local river hydrodynamics) which influence the movement of plastic debris by river water [12,27,53]. Anthropogenic modification of river hydrodynamics (e.g., by hydropeaking) and of the potential for erosion of the storage zone (e.g., by reinforcement of river banks and floodplain disconnection by flood embankments) influences transport of macroplastic. Bank reinforcement, groynes and flood embankments may disconnect a given part of macroplastic storage zone from river processes and prevent remobilization of macroplastic ( Figure 1B), similarly as it was previously found for heavy metal-polluted floodplain sediments [e.g., 54]. Flow hydrodynamics are suggested as an important control on the distance and intensity of macroplastic transport [53]. The lack of natural or artificial obstructions to river flow will probably favour the transport of plastic debris in narrow, regulated channels with simplified morphology and lacking vegetated islands and in-channel vegetation, similarly as it was previously found for large wood pieces [cf. 44]. During the transport phase, plastic debris interacts mechanically with water and sediments transported by the river, which causes that its particles become mechanically fragmented [12,22]. Their size decreases and surface area increases, implying a higher potential for their further transport as well as biochemical degradation (e.g., during the storage phase) [see 12]. The rate of mechanical degradation of macroplastic will depend on the type of plastic, number of transport and storage events, during which plastic items may interact mechanically with water and sediment load transported by river, and on characteristics of the river sediments themselves (e.g., grain size) [12]. It may be hypothesized that in high-energy mountain rivers, which transport coarse sediments, the rate of mechanical degradation of plastic will be higher than in low-energy rivers transporting fine sediments. The process of riverine macroplastic fragmentation is important because it is assumed to be a main source of microplastic in rivers [see 55] ( Figure 1A). Different densities and buoyancy capabilities of specific types of plastic debris determine the mode of its transport as bedload or suspended load [19].
Up to date, the transport of macroplastic debris has been quantified using detection of plastic particles on aerial photos taken by a drone from low heights (5 and 15 m; [35]). It was also measured by visual counting from bridges, video approaches, trackers, net sampling or from infrastructure or riverbanks [e.g., 13,18,26,27,40,56]. In future works, more details may be determined using a combination of visual counting method with hydrodynamic modelling [57]. Some insight into macroplastic transport distance may be also gained by modifications of the existing hydrodynamic models used for simulations of the transport of large woody debris [e.g., 44,58] as well as by long-term observations of plastic accumulation on riverbanks [59].

Macroplastic storage
A storage phase occurs after natural or artificial placing of plastic debris in the storage zone or when plastic debris became deposited here on the surface of river sediments or in vegetation (surface storage) or accumulated in river sediments below the river bed or floodplain surface (subsurface storage; Figure 1B). The zone of potential macroplastic storage can be assumed to be the areal extent of maximum floodplain inundation since the 1960s ( Figure 1B), because only in this zone plastic debris might have been transported and deposited by river flow. The storage occurrence depends mainly on flood flows, which transport high amounts of macroplastic debris from upstream parts of a river catchment, and on local characteristics of the fluvial system, which control its deposition. During a flood, macroplastic debris placed artificially or naturally on a floodplain surface may be transported downstream [e.g., 23,27] or covered with mineral and organic sediments and become stored without previous transport or remobilization ( Figure 1A). The duration of macroplastic storage occurring near an active channel should be shorter than in distal parts of the river floodplain. High rates of the deposition of mineral and organic sediments in the storage zone would be a factor increasing the residence time of stored plastic. It can be hypothesized that a high sedimentation rate together with low flow energy at a given location will facilitate covering of surface-stored macroplastic with mineral and organic sediments and its transformation into subsurface storage, after which its remobilization potential would decrease. The density and buoyancy of specific types of plastic debris influence the mode of their transport [19]. Seasonal changes of riparian vegetation and modification of its floristic composition may increase surface roughness, favouring plastic storage during floods. In this perspective, the timing of floods in relation to the vegetation development phase may be an important temporal control on macroplastic storage. Anthropogenic or natural modifications of riparian vegetation, which change floodplain roughness [e.g., 60], will modify the potential for macroplastic storage. Anthropogenic modifications 7 of 14 of floodplains (e.g., by dam or embankments construction) that occurred before and after the 1960s are crucial for the spatial pattern of macroplastic storage zone. For instance, flood embankments constructed after the 1960s confine an active storage zone and disconnect parts of the former storage zone, from which macroplastic cannot be remobilized during floods ( Figure 1B). The storage zone can be also narrowed downstream from dams, where the channel typically incises and the floodplain zone is constricted laterally. Upstream from dams, the lateral extent of the storage zone should widen because of a wider extent of floodplain inundation during floods induced by reservoir backwater effect and bed aggradation [61].
The amount of stored macroplastic can be calculated using data on the volume of storage zone (based on known areal extent and mean thickness of sediments) and plastic abundance in the sediments. The works on surface and subsurface storage of macroplastic debris used hand collection and sieve analysis for estimation of plastic abundance (Table 1). Analysed samples were collected from quadrats [30,34,62,63], circles [64] or transects [65] delimited parallel or perpendicular to the shoreline of river channel or reservoir [59]. Obtained results were usually described as number of plastic debris items or their mass per surface area (m 2 ) of sampled plot (Table 1), which is useful for estimation of plastic abundance in contemporary surface sediments, but does not allow for reliable comparison with volumetric samples of sediments taken from cores or river banks. Importantly, this approach also does not allow for calculation of the amount of macroplastic debris that can be remobilized from a given volume of eroded sediments. These problems can be resolved by collecting samples in a way allowing for establishing the mass of plastic in a given sediment volume [see 66]. The volumetric approach was previously recommended for comparable sampling of microplastic [14,67]. Calculation of the abundance of macroplastic debris in sediment volume should be also more useful for future works utilizing samples taken from sediment cores or river banks. The volumetric method of sample collection may be implemented to existing approaches [e.g., 62,64] giving an opportunity to study the abundance of plastic debris in deposits of contemporary fluvial forms (e.g., channel bar, floodplain, island) and to compare it with similar older deposits. Such an approach could also allow for estimating the total amount of plastic stored in a given river section or a whole river system, which is of crucial importance for reliable estimation of the potential for its future remobilization by lateral migration of river channel. New insights into the storage phase may be also gained by adaption of some field experiments developed by fluvial geomorphologists for studies on seeds deposition [e.g., 68] and approaches used up to now to determine the rate of macroplastic deposition on beaches [e.g., 34] or in tidal marshes [30]. A recent remote sensing experiment suggests that storage of larger particles of macroplastic debris (> 2.5 cm) can be effectively quantified from aerial photos taken from low heights (5 to 40 m [35,37,39]), which seems to be an effective alternative for hand collection of surface-deposited plastic debris in non-vegetated parts of river sediments or hardly accessible areas of dam reservoir shorelines or deltas.
During the storage phase plastic particles are degraded by biochemical processes (Figure 1A). From a long-term perspective, a prolonged storage phase should decrease the amount of stored and remobilized macroplastic, but produce some amount of microplastic which may be released to river waters or groundwater ( Figure 1A).

Macroplastic remobilization
Remobilization occurs when stored macroplastic debris is entrained by river flow or erosion processes (e.g., by bank erosion) and starts to be transported. Remobilization can encompass both surface and subsurface-stored macroplastic debris ( Figure 1B). Remobilization of surface-stored macroplastic occurs when the debris is entrained by flow from the surface of river bed, banks and floodplain or from riparian vegetation. Remobilization from subsurface storage can occur simultaneously, for instance, during bank erosion, when both surface-and subsurface-stored plastic particles are eroded ( Figure 1B). A timescale of the transport and remobilization phases in a given part of fluvial system is controlled by number and intensity of transport and remobilization events Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 9 July 2020 doi:10.20944/preprints202007.0169.v1 and complexity of a given system. These factors control both the potential for long-distance transport of macroplastic debris [e.g., 51] and the storage time. Tramoy et al. [21] suggest that within biophysically complex systems and river sections with high storage potential (e.g., river estuary), a remobilization-storage cycle may last for centuries. New information on these processes may be gained from numerical simulations of a transport-storage-remobilization loop for longer time spans and with different scenarios of new plastic inputs. The extent of a remobilization zone can be mapped in future works as the area of the storage zone influenced by fluvial processes in a given period. The areal extent of the remobilization zone may be calculated using information on bank erosion or channel migration rates in a given river reach. Such analysis may be done using GIS techniques and remote sensing data (e.g., aerial photos) commonly used in fluvial geomorphology to study channel changes. The amount of remobilized macroplastic may be estimated using data on the abundance of surface-and subsurface-stored macroplastic and either the area of inundated floodplain or the volume of eroded floodplain sediments. Remobilization can be expressed as the amount of macroplastic (e.g., kg or items number) remobilized from unit of storage zone (e.g., km of river length, km 2 of the area of storage zone, m 3 of the volume of storage zone) per time. Similar to macroplastic transport, future insights into macroplastic remobilization can be gained by simulations with use of existing hydrodynamic models [e.g., 44]. However, such works will need further studies of the flow parameters needed to entrain particles of plastic debris deposited on given sediments or anchored on given vegetation types [cf. 30].

Macroplastic output
The output phase is defined as natural transport of macroplastic debris out of a fluvial system (e.g., from river to the ocean or estuary) or its given part (e.g., subcatchment, section or reach of river) or its artificial removal by cleanup actions or infrastructure (Figure 1). This phase is controlled by both natural (e.g., the number and intensity of transport and remobilization events, complexity of a given river system [e.g., 21]) and artificial factors (e.g., cleanup activities [15,64]). Similarly as for transport phase, a lower complexity of river systems may favour the occurrence of macroplastic output because of a lower possibility of multiple cycles of transport-storage-remobilization to occur [cf. 21]. It can be also hypothesized that for channelized and embanked rivers (with constricted channel and floodplain zones, increased channel slope and high floodwater energy), the output potential will be higher than for naturally complex rivers with wider channel and floodplain zones and more diverse morphology and riparian vegetation, which favour storage of macroplastic and prevent its downstream transport. In turn, river estuaries were indicated as possible temporary sinks for macroplastic that may delay its input from rivers to the ocean [21,69]. Similar to transport phase, an output phase may be also temporarily modified by hydropeaking or sediment flushing events occurring downstream from dams. The intensity of an output phase may be measured in the same way as for transport phase [e.g., 27,40,56]. beach (repeated twice at the same location).

Perspectives on future work
Research on macroplastic in the river environment is an emerging branch of river science and its methodological and theoretical background is developing now. In this paper we aimed to shed light on the processes of riverine macroplastic storage and remobilization which were overlooked in the literature. Our conceptual model provides a framework that defines phases of the macroplastic route through a fluvial system and systematizes their main controls. The proposed model is based on the assumption that the fate of macroplastic in a fluvial system is controlled by natural and anthropogenic factors, the importance of which changes between different phases of its route through the system (Figures 1-2). The key factors controlling macroplastic storage and remobilization are related to natural characteristics of fluvial systems (especially river flow and valley morphology) and their anthropogenic modifications. Types and number of these controls change along different spatial units of a fluvial system and the control factors operating at larger units (e.g., catchment) additionally influence those active at smaller units (e.g., reach) ( Figure 2). The proposed model presents a general qualitative scheme, which may be quantitatively developed in future works based on a greater number of case studies. The temporal coincidence of the beginning of plastic debris accumulation in the environment in the 1960s [1] and of widespread availability of historical remote sensing materials (e.g., aerial photos) and hydrological data since that time gives a unique opportunity to map macroplastic storage and remobilization zones since the beginning of the plastic era. The information about the abundance of plastic in contemporary and older sediments of storage zone should be expressed in the same metrics, which would allow for easier comparison of the abundance of plastic debris between fluvial forms and sedimentary facies of different age and for precise assessment of the total amount of plastic stored in the river environment.
A demonstrated diversity of factors controlling the route of macroplastic through fluvial system requires a broader, transdisciplinary perspective that would include among functional components humans, who not only dispose plastic, but are also affected by it both physically and aesthetically, and who may remove it from rivers.