Plastic size and shape relationships measured in the Meuse and Rhine

A field study of two Dutch rivers found strong positive relationships among the length, measured surface area and circumference of individual plastic particles, while the number of particles per size class did not consistently decline as particle length increased. In the Meuse, where complete concentration data were available, cumulative surface-area concentration increased significantly with particle size at both sampling locations. The researchers emphasize that these site-specific results, based on the Meuse and Rhine in the Netherlands, do not establish a universal pattern for rivers and cannot yet be generalized to other river systems.

The investigation involved 318 sampling events at four locations in the Meuse and Rhine river systems. The Meuse was sampled at Eijsden and Sambeek in April and May of 2023 and 2024, while the Rhine system was sampled at Lobith and Tiel in 2023. Lobith and Tiel were treated collectively as the Rhine because the Rhine bifurcates downstream of Lobith and the Tiel site lies in the Waal, its main branch. Researchers collected material with a trawl net measuring 4 by 1 metre and having a mesh size of 6 millimetres. Each haul lasted approximately 40 minutes and sampled the surface, middle or bottom of the water column, depending on water depth and bottom obstacles.

All samples were cleaned aboard the research vessel, where visible organic material was removed. Plastic pieces were subsequently arranged on A4 paper and scanned, flattening them with paperclips when necessary. ImageJ was used to measure the longest axis, treated as particle length, as well as two-dimensional surface area and circumference. Particles too large to scan intact were cut to fit on the paper; for those pieces, surface area could be measured, but length and circumference could not. Based on their longest-axis measurements, particles were classified as mesoplastic when they measured 5 to 25 millimetres and as macroplastic when they exceeded 25 millimetres. The primary focus was on these larger fractions rather than microplastics, although some smaller pieces were recovered.

Because the mesh could retain particles smaller than its nominal 6-millimetre opening, the researchers applied an 8.5-millimetre minimum-length threshold, derived from the diagonal of the mesh opening. This produced an analytical dataset of 9,096 particles from the 9,545 pieces collected. Of the particles retained for analysis, 45.9 percent were mesoplastic and 54.1 percent were macroplastic. A second analysis using the full dataset, with two anomalous observations excluded, largely reproduced the patterns found with the threshold, including relationships involving smaller particles.

Sampled water volume was calculated from the duration of each net haul, flow velocity measured with an acoustic Doppler current profiler and the area of the trawl opening. Particle concentration was expressed as the number of pieces per 1,000 cubic metres of sampled water. Cumulative surface-area concentration was calculated differently: the measured areas of the particles in each sample were added together, then divided by the sampled volume, producing a value in square millimetres per cubic metre. These two measures represent different aspects of the recovered material, with the second accounting for the greater area contributed by individual pieces as their dimensions increase.

Across the full dataset, the three measured dimensions of individual particles were strongly and positively related. The relationship between area and length followed a power law, but its exponent was below the value of two expected if particle shape remained geometrically similar as size changed. The reduced major-axis estimates were 1.66 in the Rhine and 1.81 in the Meuse, while ordinary least-squares estimates were 1.43 and 1.67, respectively. The lower exponents indicated that relative shape changed with size, with longer particles tending to become more elongated or one-dimensional. This change was more pronounced in the Rhine, where the area-length exponents were farther below two.

Circumference followed a different scaling pattern. A theoretical relationship in which circumference is proportional to length has an exponent of one, and the observed estimates were close to that value. Across rivers and locations, reduced major-axis regression produced an exponent of 1.06, while ordinary least-squares regression produced 1.02. The researchers also examined how circumference changed with area, finding exponents below two in each river and at both upstream and downstream locations. Reduced major-axis and ordinary least-squares methods yielded different estimates in several comparisons, and their 95 percent confidence intervals rarely overlapped, even though their coefficients of determination were similar.

Particle dimensions also differed spatially. In the Meuse, particle length, surface area and circumference were all significantly larger at the upstream site than at the downstream site. The Rhine showed a different pattern: particle length was greater downstream, whereas area and circumference did not differ significantly in the same direction. The researchers could not interpret these contrasts as direct evidence of fragmentation along either river. Upstream and downstream sampling occurred on different days, and the schedule did not account for the travel time of the water, meaning that the samples did not necessarily represent the same moving body of water.

The size-abundance analysis produced results unlike the consistently inverse relationships reported in previous ecological size-spectrum studies. Ten equal-sized logarithmic length bins were created for the Meuse. At the upstream site, particle concentration generally increased with length: the reduced major-axis exponent was 1.52 and the ordinary least-squares exponent was 1.17. At the downstream site, concentration did not vary significantly across the bins, with corresponding exponents of -0.20 and -0.07. Upstream size bins also had significantly higher particle concentrations than downstream bins. Within the sampled Meuse data, therefore, particle size alone did not explain abundance in the way it did in earlier marine and ecological studies.

Cumulative surface-area concentration nevertheless increased significantly with particle length at both Meuse locations. Upstream, the reduced major-axis and ordinary least-squares exponents were 9.39 and 9.07; downstream, they were 9.29 and 6.82. The steep relationships reflect the combined contribution of particle abundance and the rapidly increasing area of individual pieces. The upstream site also had a higher cumulative surface-area concentration across the size bins than the downstream site. Because time constraints prevented complete analysis of all particles collected in the Rhine and Waal, these concentration and cumulative-area relationships could be evaluated only for the Meuse and should not be portrayed as a finding for both rivers.

Several features of the method limit the conclusions. The 6-millimetre net was designed primarily to recover mesoplastic and macroplastic debris, not the full microplastic range, and some smaller particles may have been retained after aggregating with organic matter. The two-dimensional scanning method also underestimated the true surface of three-dimensional objects. In the Meuse analysis, the bins at the beginning and end of the measured length range contained fewer observations than those in the middle, increasing the potential influence of outliers. Discharge may also have affected concentrations, but its relationship with particle abundance was not established because the sampling did not systematically cover the range of flow conditions needed to estimate a site-specific relationship.

The findings provide a more detailed starting point for river-pollution risk assessment because counts alone do not describe the dimensions of the material present. Two samples with similar particle concentrations can differ substantially in cumulative surface area when their pieces have different sizes or shapes. The researchers recommend broader monitoring across a wider size range, repeated sampling of the same water parcels to investigate fragmentation, measurements of particle mass and density, and additional observations under different discharge conditions. The study characterizes environmental concentrations and scaling relationships; it does not directly measure ingestion by organisms, chemical exposure, ecological harm or human health effects, and its concentration results remain specific to one Dutch river.

Subject of Research: Measurements of plastic particle dimensions, abundance and cumulative surface area in the Meuse and Rhine rivers

Article Title: Scaling of dimensions and concentrations of plastics in two Dutch rivers

Article References: Polinder, J., Collas, F. P. L., Leiva, F. P., Oswald, S. B., Schoor, M. M., Verberk, W. C. E. P., & Hendriks, A. J. (2026). Scaling of dimensions and concentrations of plastics in two Dutch rivers. Environmental Monitoring and Assessment, 198(11), Article 1227. https://doi.org/10.1007/s10661-026-16029-y

Image Credits: AI Generated

DOI: 10.1007/s10661-026-16029-y

Keywords: river plastic pollution, plastic particle size, particle shape, cumulative surface area, Meuse River, Rhine River, mesoplastics, macroplastics, environmental monitoring, size spectra, Scaling, dimensions