The Tisza River, which originates in the Bukovina segment of the Carpathian Mountains, is the main tributary of the middle Danube River. Its length and drainage area are 966 km and 157,186 sq. km, respectively. Along the Hungarian section of the Tisza River, four sampling sites at the settlements of Tímár, Tokaj, Csongrád, and Szeged, which are located at flow km of 552, 544, 244, 160, respectively, were selected (Figure 1). The sediment samples were collected in the time period of August 2–4, when the algal population was relatively high and defined by relatively high chlorophyll-a levels (20 μg/L). At all sampling sites, three sampling points located about 6–8 m from the river bank and spaced roughly 4–5 m apart were selected. At these sampling points, 5 L of suspension was pumped into amber bottles from the 10–15 cm thick suspension layer streaming over the bottom of the riverbed using a portable, pressure-difference “SEDIMONER” sampler developed by Aqua-Terra Lab Ltd. (Veszprém, Hungary). Although this sampling method does not produce an isokinetic sample, it is more accurate when sediments are fine (<63 µm) and flows are turbulent. Van Veen grab sampler was used for sampling bottom sediments, allowing samples to be collected from a nearly 10 cm thick surface layer of bottom sediments. About 1 kg of sample was collected at each sampling point and transferred into plastic bags with zip-fastener. The suspension and sediment samples were stored on ice during transport to the laboratory.

For the characterization of the water phase, the water temperature, pH, dissolved oxygen concentration, redox potential, electric conductivity, and turbidity were measured on-site using a portable Ponsel-ODEON water quality meter (Fondriest Environmental, Ohio, United States). During the campaign, the flow velocity was measured by a rotating current meter (Global Water Flow Probe FP 211, Xylem, United States). The concentrations of selected anions, cations, as well as total organic C and N of water samples obtained after filtration of NBS sediments were determined by applying DIONEX 5000 ICS + dual channel ionchromatograph (Thermo Scientific, Massachusetts, United States) and a MULTI N/C 3100 TC/TN analyzer (Analytik Jena, Jena, Germany), respectively. The elemental concentrations were quantified using an inductively coupled plasma mass spectrometer (Analytik Jena, Jena, Germany). These parameters are listed in the Supplementary material (Supplementary Tables S3, S4).

A WasserLab Automatic unit (Labsystem Ltd., Budapest, Hungary) was used for generating ultrapure water (resistivity of 18.2 MΩ cm⁻¹). Nitric, hydrochloric, and sulphuric acids, hydrogen peroxide, and solid natrium hydroxide of analytical grade were purchased from VWR International Ltd. (Debrecen, Hungary). The internal and multi-elemental standard solutions were produced by Sigma-Aldrich Ltd. (Missouri, United States).

The suspension samples were homogenized and three samples of 200 mL each were separately filtered using pre-combusted glass fiber filters with a pore size of 0.7 µm. The empty filters were weighed before the filtration. The loaded filters were dried at 80°C for 12 h and re-weighed to determine the dry mass of solid particles for calculation of the mass concentration of TSS.

The bottom sediments were wet-sieved over a 2 mm mesh to eliminate large detritus and benthic organisms and dried at 80°C for 12 h in a laboratory oven together with the suspended sediments. Three replicates of both sediment samples were individually homogenized. About 30–90 mg of these samples were resuspended in 200 cm³ ultrapure water and subjected to 5 min of ultrasonic treatment at a frequency of 32 kHz. The cumulative and differential distribution functions were determined with the use of a Shimadzu SALD-2300 laser diffraction particle size analyzer (Shimadzu, Kyoto, Japan).

For measuring the TOC contents, 300–500 mg dried and homogenized sediment samples were analyzed using a Multi N/C 3100 analyzer (Analytik Jena, Jena, Germany). The total Kjeldahl nitrogen (TN) of dried sediments was measured by automated colorimetry with preliminary distillation/digestion on the basis of Standard Method 1,687 (EPA-821-R-01-004, January 2001).

To determine the pseudo-total concentration of phosphorus (TP), arsenic, and different metals, 500 mg dried sediments were treated with 8 cm³ aqua regia at a temperature of 200°C for 20 min in a TopWave microwave-assisted digestion system (Analytik Jena, Jena, Germany). Three replicates were prepared for all sediment samples. Following the sedimentation of solid particles (predominantly silicates), 2 cm³ of the clear solution was removed by a pipette and diluted 25-fold with ultrapure water. After the addition of internal standards (Sc, Y, In) in concentration of 20 μg/L, the main (Al, Fe, Mn, P) and trace elements (As, Cd, Co, Cr, Cu, Hg, Li, Ni, Pb, Sn, Zn) were determined using a Plasma Quant Elite inductively coupled plasma mass spectrometer (Analytik Jena, Jena, Germany). The operating conditions are listed in Supplementary Table S5. To characterize the reliability of this analytical procedure a recovery test was carried out for 8 elements (Cd, Cr, Cu, Hg, Mn, Ni, Pb, Zn) analyzing the BCR-146R (Sigma Aldrich, Missouri, United States) certified reference material. The recovery values changed in the range of 82% and 114%.

Data were visualized using R statistical software (R Core Team, 2019). The relationship between TOC and metal/metalloid concentrations was examined by Person’s correlation after checking the assumption of the normal distribution. Figures 3A–D were drawn with the tkplot function of the “igraph” package (Csardi and Nepusz, 2006) using the Pearson’s correlation coefficients and p-values as input data.

The mass concentration of NBS sediment particles fluctuated between 209 and 274 mg/L. These values are similar to TSS concentrations of 70–247 mg/L measured at middle depth in the Yellow River, however, considerably lower than the TSS concentrations in the 2:1:1 ratio mixed subsurface, middle and near-bed water samples of Yangtze River (Wang et al., 2012).

The grain size distribution functions of suspended and bottom sediments originating from the same sampling site are depicted in Figures 2A–D. The bottom sediments can be characterized by a high maximum value within the typical sand size range of 60–500 μm, while the NBS-sediments primarily consist of silt (size range 2–60 µm). In its grain size distribution functions, three different grain size groups can be observed, signifying the existence of particles originating from various sources. In all NBS sediments the maximum values are observable approximately at particle sizes of 6, 20, and 60 μm. Considering the particle capture efficiency of mussels, even this size range is important for the nutrient supply of Unionidae mussels (Kryger and Riisgård, 1988). However, (Tuttle-Raycraft and Ackerman, 2018), concluded in their paper that particle quality rather than size is what modulates suspension feeding in turbid waters. In the typical clay grain size range (<2 µm), particles were not detectable in the bottom sediments collected at Tokaj and Szeged. This phenomenon can be explained by the hydrodynamic sorting of the rivers (higher stream velocities at these sampling points (0.25 vs. 0.1 m/s). However, in all NBS sediment samples, around 5%–10% of the particles were found to be in this size range. In conclusion, silt predominates among the NBS sediments investigated, which is consistent with the finding made by (Bouchez et al., 2011) in the instance of the Amazon River.

It should be noted that our data on grain size measurements do not accurately reflect the grain size distributions that originally existed in the mussel habitats. This is because the cohesive suspended sediments are transported primarily as flocculated material (Droppo, 2001; Deng et al., 2019), but the flocs were damaged during the sample preparation by drying and resuspension of solid particles applying ultrasonic treatment before the grain size measurements. It implies that the effect of flocculation processes governed by extracellular polymeric substances (Lee et al., 2017) is partly eliminated. In order to avoid these undesirable disturbances and to obtain a true picture of the particle sizes of suspended sediments, it would be necessary to apply in-situ methods (e.g., submersible particle size analyzer and submersible digital holographic camera), as was demonstrated by (Safar et al., 2022). In the absence of such devices during our investigations, the suspended particles were collected and analysed in the conventional manner described in subchapters 2.1 and 2.4.