Insights into suspended sediment and microplastic budget of a lowland river: integrating in-situ measurements, Sentinel-2 imagery, and machine learning

Abstract

The calculations on fluvial microplastic load (MPL) provide dynamic and actionable metrics for understanding microplastic (MP) particle emissions to the downstream environment. However, most of the concentration data reported in the literature do not reflect the total amount of transported MPs. This study aims to quantify the MPLs in the Middle and Lower Tisza River, Hungary by combining multiscale, frequent in-situ measurements of MP concentration (MPC) and suspended sediment concentration (SSC) in the river with Sentinel-2 and ANN-based models. MPC data were integrated with daily water discharge, applying correction factors to account for vertical variability in the water column. The results indicate the crucial role of hydrology in suspended sediment (SS) and MP transport. During floods SS load (SSL) (6706 ± 7449 ton/day) and MPL (1.71 × 10⁹ ± 1.40 × 10⁹ item/day) were 6.5 and 5.1 times higher than at low stage (SSL: 1031 ± 755 ton/day; MPL: 0.339 × 10⁹ ± 0.399 × 10⁹ item/day), respectively. Vertical measurements indicated a generally increasing SSC trend toward the riverbed, amplifying the SSL correction factor, while decreasing MPC led to attenuation for MPL correction factors. Satellite-derived estimates at four sites in the Middle and Lower Tisza revealed an increasing downstream trend in SSL and MPL, although natural and anthropogenic factors slightly influence this trend. Rating curves were established at the four sites with an R² range of 0.55 (Mindszent) to 0.82 (Zenta) for SSL and 0.53 (Mindszent) for MPL. Although the accuracy of the rating curves is moderate to very strong, they still offer practical predictions based solely on water discharge data.