Coupling Particle-Scale Hydrodynamics with Landscape-Scale Spatio-Statistical Topology: A Predictive Framework for Microplastic Fate in the Sebou Estuary–Atlantic Coast System (Morocco)

The interplay between particle-specific hydrodynamics (settlement velocity, shape-specific drag, resuspension thresholds) and landscape-scale forcing (population-specific density, tidal asymmetry, sedimentological trapping, longshore drift) determines the fate of microplastics (MPs). The central challenge in the present study is the recognition that the existing literature suffers from a conceptual disconnect between the landscape-scale patterns of MPs and their physical dynamics at the particle scale: the former ignore the processes of vertical partitioning and coastal export, while the latter fail to acknowledge the heterogeneity of spatial clusters and source areas. First, we utilize published data on the abundance and composition of MPs in the Sebou Estuary and Atlantic Coast of Morocco to establish a novel conceptual–hydrodynamic–spatial model that simultaneously accounts for these processes. Published data include 18 stations with 10–298 (~300) particles kg−1 in sediments and 10–168 particles m−3 in waters sampled on 16 December 2020. Second, we explore the relative roles of size and polymer-specific density, shape-specific drag, and tidal asymmetry in the observed patterns using a modified Stokes–Ganser settling solution, tidal pumping with Shields criterion, Getis–Ord Gi clustering analysis, ecological topology, non-metric multidimensional scaling, hierarchical clustering, and 2D predictive risk mapping. Fibers (PET/nylon, ρ ≈ 1380 kg/m3) settled 3.2–4.0× faster than fragments of the same size (PE/PP, ρ ≈ 920 kg/m3) according to the settling model, explaining the finding that fibers were abundant in Sebou sediments but rarely detected in water. Granules (industrial pellets, 1–5 mm) were not found in water as they rapidly settled to the bottom (at >38 mm/s) and accumulated on the 2 m deep bed within 10–52 s. Bed shear stresses (τb) in neap tides (τb ≈ 0.18 Pa) were lower than critical resuspension stresses (τc) of all sinking MPs, meaning that the bed retained them, while spring currents (τb ≈ 2.16 Pa) were sufficiently turbulent to erode particles > 0.1 mm and re-suspend them in the water column. Consequently, their export followed a fortnightly cycle determined by the critical shear-stress analysis with the Shields criterion. According to Getis–Ord Gi analysis, Kenitra Harbor (E5: Z = 11.85, p < 0.01; E6: Z = 8.94) and Mehdia Harbor (E1: Z = 6.42; E2: Z = 5.18) were associated with four statistically significant spatial clusters of MPs that showed a similar pattern of contamination, rather than being isolated anomalies. Non-metric multidimensional scaling (stress = 0.029) confirmed the presence of three distinct contamination regimes, namely background (reference), coastal transport corridor (B7–B10), and urban-sewage (E1, E2, E5, E6), as did hierarchical clustering. The correlation network analysis identified a diagnostic film–granule anti-correlation (r = −0.72) that could be used to distinguish domestic plastics from industrial pellets without spectral analysis. Consistent with the overall pattern, 1D advection–diffusion–decay modeling of the longshore current indicated that MPs would accumulate at Oulad Berjal Beach (B10) with a northward drift of ~0.3 km/day. According to the Monte Carlo simulation of n = 10,000 iterations, a 22% variation in settling velocity would account for a 35% variation in the transport distance. The Sebou Estuary is a local microplastic hotspot with a self-sustaining contamination cycle, while the most promising options for remediation target the identified clusters of anthropogenic input.