The Trash-Tracker: A Macroplastic Transport and Fate Model at River Basin Scale

1 Hydrology and Quantitative Water Management Group, Wageningen University, Wageningen, the Netherlands 2 Department of Earth Sciences, Faculty of Geosciences, Utrecht University, Utrecht, the Netherlands 3 Aquatic Ecology and Water Quality Group, Wageningen University, Wageningen, the Netherlands 4 Climate and Environmental Physics, Physics Institute, University of Bern, Bern, Switzerland 5 Oeschger Centre for Climate Change Research, University of Bern, Bern, Switzerland 6 Department of Marine Microbiology and Biogeochemistry, NIOZ Royal Netherlands Institute for Sea Research, ‘t Horntje, the Netherlands 7 CAGE Centre for Arctic Gas Hydrate, Environment and Climate, Department of Geosciences, UiT The Arctic University of Norway, Tromsø, Norway


Introduction
Plastic pollution causes harm to wildlife (e.g. ingestion or entanglement [40]) and has negative impacts on using a fictional case study using real-world forcing data.

72
The model concept is based on a principal criterion in the field of sedimentology, which states that sediment over land is a function of driving and resistive forces as well and that thresholds mark the conditions required for 75 incipient motion (Fig. 1). The two driving forces in the model are wind ( ) and surface runoff ( ) (the same driving 76 forces were used by Meijer et al. [29]) and the resisting force, i.e. the terrain friction, is a result of the combination 77 of land use and terrain slope, which is translated to a wind ( ℎ ) and a surface runoff threshold ( In case none of the thresholds are surpassed (eq. 1), the macroplastics will not be mobilised and no transport 84 occurs. If only the wind threshold is surpassed (eq. 2), the macroplastics will move in the direction of the wind at 85 that geographic location. In case only the surface runoff threshold is surpassed (eq. 3), the macroplastics will move 86 in the direction of the surface runoff, which is equal to the direction of the steepest downhill terrain slope at that 87 geographic location. Finally, if both thresholds are surpassed (eq. 4), the model randomly picks either the wind or 88 the surface runoff direction at that geographic location along which the macroplastics will move.  Fig. 2). It is assumed that the mismanaged macroplastic waste generated during a single time step is exposed to the weather conditions of that 94 same time step and is immediately available for transport.

154
Option 2: For this calculation of the wind thresholds it was assumed that in case of winds blowing uphill/downhill, 155 the ability of the wind to mobilise and transport macroplastics in the direction of the wind decreases (uphill winds) 156 or increases (downhill winds), because it is counteracted (uphill winds) or assisted (downhill winds) by the force of 8 wind speed but in a different direction appears to be sufficient to surpass the wind speed threshold and consequently 166 moves the macroplastics.
Surface runoff driven transport such surface runoff thresholds. We made a first attempt and established the orders of magnitude for our surface runoff thresholds on the distribution of the data on global absolute runoff trends found in the Global Runoff Ghiggi et al. [13]. We assumed that the higher the density of natural (e.g. vegetation) or anthropogenic obstacles (e.g. buildings), the more surface runoff is required to displace macroplastics.   can also be computed from rainfall data (extracted from regional/national weather stations) using a runoff coefficient.

247
The runoff coefficient (= runoff / rainfall) is the fraction of the rainwater that does not infiltrate in the soil and

248
consequently becomes surface runoff. The type of land cover (i.e. vegetation) plays a major role in this process.

249
The surface runoff direction in each grid cell is equal to the direction of the steepest terrain slope of that grid cell 250 ( Supplementary Fig. SI2).

251
The rainfall values used in the model application can be found in Supplementary Fig. SI4. These values were 252 generated on the basis of a frequency

255
The surface runoff flux and direction in river grid cells equal the river flow speed and direction, respectively. The clean-up operations that prevent plastics to enter the marine environment.

327
The spatiotemporal macroplastic distribution maps generated by the model application demonstrate that densely

337
The Trash-Tracker keeps track of where the MPW is located within the river basin at all times so that, for 338 example, the ratio between MPW on land and afloat in the river can be determined. Plastic retention in the river is

392
The Trash-Tracker registers the plastic emissions for each time step and thereby allows for accurate river basin 393 specific data on where and when plastics leave the river basin. This provides valuable insights on the response 394 (time) of plastic emissions to seasonality and extreme weather conditions (e.g. floods or hurricanes) and is crucial 395 for anticipating peak plastic discharges at river mouths. But most importantly, the spatiotemporal data on the plastics 396 emissions of single river basins can be used as input for oceanic plastic particle tracking models.

397
In the model application, we found that from all the plastics that left the (hypothetical) river basin, 88.1% was 398 emitted by the river, 5.0% by the coast (bare land in Fig. 4b) and 6.9% was moved to the adjacent river basin (land- All these aspects could be included in the modelling framework, when better parametrization of the different 460 processes would become available.

461
Another aspect that is difficult to parameterize is the influence of anthropogenic structures and activities on waste items (e.g. size, shape, density, wet/dry, etc.) on the mobilisation and transport thresholds [37].

477
Once the Trash-Tracker contains empirically proven mobilisation and transport thresholds, the model predictions

482
We anticipate that future collaborations with field collection and monitoring projects allow for a fast and robust calibration of the Trash-Tracker and improve the validity of the forecasted transport and fate of macroplastics within river basins.

Conclusions
Each

729
Bare land is assumed to have the lowest resistance to wind driven plastic transport, due to the absence of natural (e.g. vegetation) and artificial (e.g. buildings) obstacles. The lowest wind speed threshold has been set to 6.6 m/s, as according to the Beaufort scale, wind speeds of BF 4 (5.8 -8.8 m/s) "raise dust and loose paper" This value has been extrapolated for higher resistance terrains types, based on plastic transport probabilities estimated by a group of 24 experts. These 24 experts were asked to answer the following questions (see Table S8 in the Supplementary Materials from Meijer et al.

[3]):
 What is the overland transport probability per kilometre for land use type 'bare land'?
 What is the overland transport probability per kilometre for land use type 'urban'?
 What is the overland transport probability per kilometre for land use type 'forest'?  (Fig. SI1). As no probability of plastic transport estimates were available for grass/shrublands, we

830
Each day was assigned to one of the 23 rainfall classes based on the total amount of rainfall that fell during that day. Subsequently, 831 the frequencies for each rainfall class were computed. We did not take monthly variations into account, therefore the frequencies 832 in the  global riverine plastic emissions into the ocean. EarthArXiv. 2019 http://dx.doi.org/10.31223/osf.io/zjgty.