This paper focuses on the Harbin section of the Songhua River in the cold region of northern China. The occurrence characteristics and pollution level of microplastics (MPs) are explored in both the ice and water of urban inland rivers and estuaries during the ice-covered periods. The abundance of MPs in Majiagou ice averaged 324.4 ± 261.5 particles/L, and the abundance of MPs in Songhua ice averaged 65.0 ± 68.2 particles/L. In the water with different depths of Songhua, the MP abundance ranged from 1.0 ± 0.7 particles/L to 12.9 ± 9.4 particles/L, with an average of 5.6 ± 7.6 particles/L. The amount of MPs in ice is about 11 times that in water, where ice formation is mainly responsible for the decline in the MP abundance in aquatic environments. The abundance of MPs in urban inland rivers gradually increased from south to north, while that in the mainstream of the Songhua River showed an increasing trend from east to west. Detected MPs were mainly fiber and white in shape and color, respectively, with a particle size < 0.5 mm. The extent of microplastics in ice is greater than that in water, and melting in the following spring will exacerbate the environmental impact. The results identified the discharge of domestic sewage as the main source of MPs in urban inland rivers. Polyethylene, polypropylene, and polyacrylonitrile were the main types of polymers. The results of the ecological risk assessment showed that the MP pollution in the Harbin section of the Songhua River reached moderate and severe pollution levels during the ice-covered periods. Its potential risk should receive more attention, and control should be strengthened.

Polyethylene microplastics (PE MPs) occur widely in soils, but how particle size and concentration relate to earthworm physiological responses and gut bacterial communities remains unclear. We exposed Eisenia fetida to four nominal PE particle sizes (15, 45, 100, and 200 μm) at three concentrations (0.5%, 2%, and 5%, w/w) for 30 days. Mortality, container-level mean body-weight change, cocoon production, gut bacterial communities, and PICRUSt2-predicted functional potential were assessed. Mean body-weight change was lower in all 12 PE MP treatments than in the control (Dunnett-adjusted p < 0.001), with the most negative mean value (−19.50 ± 3.80%) observed at 15 μm and 5%. Particle size and concentration had significant main effects on body-weight change, but no significant interaction. Mortality and cocoon production did not differ significantly from the control after adjustment. Neither Chao1 richness nor Shannon diversity differed significantly among groups. Global bacterial community differences were detected (PERMANOVA, R² = 0.393, p = 0.0001), although unequal dispersion (PERMDISP, p = 0.0069) and nonsignificant BH-adjusted pairwise comparisons limited their interpretation. None of the ten dominant phyla or genera tested differed significantly among groups after false-discovery-rate correction, while LEfSe yielded exploratory discriminatory taxa. Only three KEGG Level 2 and six Level 3 treatment–control comparisons met the BH-adjusted significance threshold within their respective comparison families. Overall, container-level body-weight change showed the most consistent treatment–control differences, whereas evidence for treatment-specific gut bacterial and predicted functional changes was more limited under these high-exposure laboratory conditions.