The residence times of the HD black pellet tracers in each monitor box of the Nazaré Canyon model simulation in (a) the spring/summer 2009 model run and (b) the autumn/winter 2011 model run. 

Contexts in source publication

Context 1

... the model output during both spring/summer and au- tumn/winter periods, pellet tracers throughout the canyon travelled greater distances than they were displaced (Fig. 2), further indicating the tracers were transported in an oscillat- ing manner, up and down canyon repeatedly, as suggested by the residence times shown in Fig. ...

Context 2

... the spring/summer period, boxes 1 and 3 exhibited similarly low displacements, distances and velocities of pel- let tracer transport (Fig. 2), but had the largest difference in the percentage of tracers that ultimately escaped the moni- tor box (Table 4). Similarly, transport behavior of pellet trac- ers in box 2 and box 4 was similar, but the percentage of tracers escaping from monitor box 4 was a factor of 3 times higher than that modelled for box 2. These incongruences suggest that pellet movement within and pellet export from the monitor boxes are not necessarily correlated. Alterna- tively to the spring/summer period, during the autumn/winter period, pellet tracers in boxes 1, 2, and 4 exhibited consis- tently high distance, displacement and velocity behaviors as compared to pellet tracers in box 3 (Fig. 2) but overall, the percentage of escaped tracers increased with depth in the au- tumn/winter modelled run, and, with the exception of box 3, in the spring/summer modelled run (Fig. 3, Table 4). ...

Context 3

... residence times of the HD black pellets in each moni- tor box are depicted in Fig. 3 as the fraction of pellet trac- ers remaining inside the monitor box over time. In all cases of both the spring/summer and autumn/winter model runs, ∼ 90-100 % of the pellet tracers were not transported outside Table 3. The mean Feret's diameter and density -ρ of HD black pellets from the Los Angeles beach sample. Settling velocity, bed load -τ b , critical -τ cr and depositional -τ d shear stress of the HD black pellets as tested in saltwater of ρ = 1026.69 kg m −3 . Mean diameter -d 50 , density, settling velocity, critical and depositional shear stresses of BBL aggregates (Thomsen et al., 2002;) for comparison to HD black pellets. Standard deviations indicate differences between measurement replicates. their monitor box at the culmination of the 101-day model run period. Extrapolating these data, using an average of 95 % of pellet tracers remaining at the end of 101 days al- lows the average monitor box residence time for microplas- tics in this study to be estimated at ∼ 5.5 years. The minimal removal of tracers from monitor box 1 indicates that resi- dence times close to shore may be on the decadal scale, as- suming hydrodynamic conditions to be similar to those mod- elled for the entire duration. In contrast, for pellet tracers from monitor box 4 (from which > 5 % of the tracers were modelled to be transported during the 3 month autumn/winter period modelled), the residence time can be estimated to be shorter, at ∼ 3.5 years. Integrating over the entire 200 km- long canyon system ( Tyler et al., 2009), and assuming rela- tively constant hydrodynamic conditions, the results suggest benthic microplastic transport from the canyon head to the abyssal plain would take place on the centennial or longer scale, if at all. Two signature patterns can be observed in the residence time plots; first, the regular sinusoidal oscillations of the pel- let tracers and second, the irregular occurrence of distinct transport events where the fraction of pellets in a monitor- ing box decreases or increases abruptly (Fig. 3). The reg- ular pellet fraction fluctuations correspond with the diur- nal and semi-diurnal component of the tide observed in the Iberian margin region. These sinusoidal fluctuations indica- tive of tidal forces are observed at all depths for the dura- tion of both simulation periods, but are particularly evident in the shallower monitor boxes 1 and 2 and decrease consis- tently in amplitude with depth. However, regardless of depth, the tidal water movement appears to have very little effect on modelled net displacement of pellet tracers. Rather, sud- den changes (primarily decreases) in the fraction of tracers in monitor boxes suggest transport occurs primarily within the canyon during occasional periods of elevated oceano- graphic flow conditions, e.g. sediment gravity flows or in- ternal waves which are reported to occur in the canyon (De Stigter et al., 2007;Quaresma et al., 2007;Muacho et al., 2013). These sudden transport events often occurred simulta- neously in each box, signifying that these pellet tracer move- ments were likely the result of a large-scale event, rather than small-scale surface ...

Context 4

... residence times of the HD black pellets in each moni- tor box are depicted in Fig. 3 as the fraction of pellet trac- ers remaining inside the monitor box over time. In all cases of both the spring/summer and autumn/winter model runs, ∼ 90-100 % of the pellet tracers were not transported outside Table 3. The mean Feret's diameter and density -ρ of HD black pellets from the Los Angeles beach sample. Settling velocity, bed load -τ b , critical -τ cr and depositional -τ d shear stress of the HD black pellets as tested in saltwater of ρ = 1026.69 kg m −3 . Mean diameter -d 50 , density, settling velocity, critical and depositional shear stresses of BBL aggregates (Thomsen et al., 2002;) for comparison to HD black pellets. Standard deviations indicate differences between measurement replicates. their monitor box at the culmination of the 101-day model run period. Extrapolating these data, using an average of 95 % of pellet tracers remaining at the end of 101 days al- lows the average monitor box residence time for microplas- tics in this study to be estimated at ∼ 5.5 years. The minimal removal of tracers from monitor box 1 indicates that resi- dence times close to shore may be on the decadal scale, as- suming hydrodynamic conditions to be similar to those mod- elled for the entire duration. In contrast, for pellet tracers from monitor box 4 (from which > 5 % of the tracers were modelled to be transported during the 3 month autumn/winter period modelled), the residence time can be estimated to be shorter, at ∼ 3.5 years. Integrating over the entire 200 km- long canyon system ( Tyler et al., 2009), and assuming rela- tively constant hydrodynamic conditions, the results suggest benthic microplastic transport from the canyon head to the abyssal plain would take place on the centennial or longer scale, if at all. Two signature patterns can be observed in the residence time plots; first, the regular sinusoidal oscillations of the pel- let tracers and second, the irregular occurrence of distinct transport events where the fraction of pellets in a monitor- ing box decreases or increases abruptly (Fig. 3). The reg- ular pellet fraction fluctuations correspond with the diur- nal and semi-diurnal component of the tide observed in the Iberian margin region. These sinusoidal fluctuations indica- tive of tidal forces are observed at all depths for the dura- tion of both simulation periods, but are particularly evident in the shallower monitor boxes 1 and 2 and decrease consis- tently in amplitude with depth. However, regardless of depth, the tidal water movement appears to have very little effect on modelled net displacement of pellet tracers. Rather, sud- den changes (primarily decreases) in the fraction of tracers in monitor boxes suggest transport occurs primarily within the canyon during occasional periods of elevated oceano- graphic flow conditions, e.g. sediment gravity flows or in- ternal waves which are reported to occur in the canyon (De Stigter et al., 2007;Quaresma et al., 2007;Muacho et al., 2013). These sudden transport events often occurred simulta- neously in each box, signifying that these pellet tracer move- ments were likely the result of a large-scale event, rather than small-scale surface ...

Context 5

... model simulations presented here indicate that the dispersion of ∼ 5 mm diameter HD microplastic pellets through the Nazaré Canyon during typical spring/summer season conditions is likely slow, but may increase with the Fig. 8. The critical bed shear stress erosion curve for quartz relates particle (sediment) size to critical shear stress, τ cr , and includes average diameter (d 50 ) 4000 µm benthic boundary layer aggregate data point (Thomsen et al., 2002). The mean HD black pellet size (d 50 ∼ 4700 µm) and τ cr is plotted over the curve for comparison of aggregate and plastic erosional behavior. intensification of the hydrographic regime during the au- tumn/winter season. The model output suggests that topogra- phy restricts pellet movement at the head of the canyon mak- ing it a potential accumulation area for non-buoyant plastic debris, but allows escaped pellets to disperse more quickly at the shelf break and in deeper sections with the occurrence of large-scale hydrodynamic events due to the widening and deepening of the canyon axis. The canyon, situated on the western Iberian Margin of the eastern North Atlantic Ocean, is a hydrographic region characterized mainly by tidal cur- rents, internal waves, and upwelling ( Vitorino et al., 2002;Quaresma et al., 2007). Throughout the upper canyon, pellet movement appears from the model to be consistently affected by tidal forces as can be seen in Fig. 3; with predicted resi- dence time of pellets fluctuating in a sinusoidal pattern char- acteristic of the peaks of the M 2 tide. Recirculation of tidal currents within the canyon has been shown to actively resus- pend organic and fine-grained lithogenic material (de Jesus De Stigter et al., 2007). From our model, it appears that tidal bottom shear stresses may also be sufficiently high, not for resuspension, but for bed load transport of HD plastic pellets in the upper canyon and to a lesser degree in the lower canyon. Near the shelf break, (boxes 2 and 3 at ∼ 300 m) where the canyon axis begins to deepen and widen (Fig. 7), the bottleneck shape (Fig. 1) of the canyon may be a source of higher current veloci- ties and wave-induced turbulence, resulting in increased pel- let transport, down-canyon. Tidally induced internal waves moving shoreward may amplify as they move into the shal- low, narrow canyon head, causing them to destabilize at the shelf break (Fig. 4), generating turbulence and elevated bot- tom shear stresses sufficient to resuspend microplastic debris on the seafloor. The regular and synchronized occurrence of sudden transport events in the three deepest monitoring boxes suggests that these events correlate with a regular hydrody- namic event, such as tidally induced internal waves, with as- sociated sediment gravity flows and resuspension of bottom material. The variable effect of such events on pellet trans- port in each monitor box indicates that the local canyon to- pography depth also affects the degree to which the large- scale hydrodynamic disturbance can transport pellets down- canyon. Other similar phenomena, such as dense water cas- cades and turbidity flows, which can occur within canyon systems due to large-scale hydrodynamic processes, are not considered here as they are not known to occur in the Nazaré Canyon on yearly timescales (De Stigter et al., 2007). The model run results support the hypothesis that in certain areas of the canyon internal waves are likely significant contribu- tors to resuspension and mobilization of deposited materials, such as ...

Context 6

... exceptionally strong transport event can be seen in Fig. 3b, where 10 % of pellets were transported out of box 2 on 1 October 2011 of the autumn/winter model run. To analyze the cause of this strong dispersion event, which was identified to occur between 18:00 and 19:00 UTC, we investigated the model input data for several days preceding the event. As wind shear stress could influence pellet trans- port at 262 m depth, wind patterns were analyzed but these did not show any significant change in direction or inten- sity. River discharges have been reported to influence current regimes on the western Iberian margin ( Oliveira et al., 2007;Martín et al., 2011); however, flow data provided by the hy- drometric stations (www.snirh.pt) from the Douro, Mondego and Tagus rivers near the canyon head showed that the con- tribution from those rivers' discharges was not significant during this period. Lastly, the modelled bottom currents in box 2 were analyzed and these showed a consistent pattern in the area during this period, with peak horizontal and verti- cal velocities in the bottom layer. The horizontal and vertical velocity modulus along the canyon axis for the time period between 17:00 and 20:00 UTC on 1 October 2011 is plotted in Fig. 4. Velocities up to 0.3 m s −1 were modelled to oc- cur in the bottom layer in box 2, which are sufficient to sus- pend and transport pellets. Bottom shear velocities near box 2 during this time period were a factor of 3 times higher than those used to suspend HD black pellets during laboratory experimentation (Tables 1 and 3). Considering water den- sity (ρ = 1026.69 kg m −3 ) used in laboratory simulations are similar to modelled water density at box 2 (σ T ∼ 27 kg m −3 ) the shear stresses experienced by the pellets in the model should be comparable to laboratory experimentation. To de- termine whether or not the increased current velocity was an artifact of an internal wave, we plotted the isopycnal surfaces, σ T (kg m −3 ), throughout the water column and along the axis of the canyon, arraying a series of plots for the same time pe- riod on 1 October 2011 (Fig. 5). The propagation of an inter- nal wave with amplitude up to 200 m is visible approaching and breaking within the upper canyon. To further verify that the pellet transport was caused by the passing of the internal wave, we compared the mean isopycnal surfaces and mean velocity modulus along the canyon axis between the 1 Oc- tober and a day during which pellet transport was modelled not to occur: 19 October 2011 (Fig. 6). The increased bottom shear stresses near box 2 at the shelf break coincide with the development of an internal wave at the head of the canyon ( Fig.s 4 and 5). Investigation of the bottom current velocity and isopycnal surfaces on the 1 and 19 October 2011 suggest that mean current velocities are insufficient to induce uni- directional transport of pellets, but amplified current veloci- ties due to large-scale hydrodynamic events, such as internal waves, are ...

Context 7

... mentioned in Sect. 3.1, abrupt displacements of pellet tracers from monitoring boxes were more evident in the resi- dence time plots (Fig. 3) during the autumn/winter model run than in the spring/summer run. The other modelled transport parameters (distance, displacement and velocity) also dif- fered by season, particularly for the shallow monitor boxes. Distance and displacement values were consistently higher during autumn/winter, up to factors of 5 or 6 times higher for pellet tracers originating in boxes 1 and 2 (Fig. 2). The standard deviations in these values were consistently higher during the autumn/winter period, as were the maximum mod- elled values, a trend indicative of a more chaotic and variable hydrodynamic environment (Table 4). Average pellet tracer velocities ranged between ∼ 0.1 and 1.0 km yr −1 in the spring/summer period and ∼ 0.2 to 0.9 km yr −1 in the autumn/winter period suggesting that pel- let velocity did not change significantly between seasons (Table 4). However, the average tracer velocity in the au- tumn/winter (0.58 km yr −1 ) was slightly greater than in the spring/summer (0.44 km yr −1 ). As with the other modelled parameters, the maximum values and standard deviations making up this average were also higher during the au- tumn/winter period (Table ...

Context 8

... the model runs, large-scale transport events appear to occur more regularly and with similar impact in each mon- itor box in the spring/summer period as compared to the autumn/winter period. During the autumn/winter, when these sudden events were modelled to occur, they impacted pel- let transport to different degrees in each monitoring box. For example, the large transport event modelled to occur on day 30 of the autumn/winter run effectively removed ∼ 10 % of pellet tracers from box 2, but had only a minimal effect on concentrations within boxes 1, 3 and 4 (Fig. 3). ...