Microplastics in wetlands: contrasting fish contamination between mangroves and temporary ponds in southeastern Brazil

<title>Abstract</title> Microplastic pollution is ubiquitous in aquatic ecosystems, but comparative analyses across wetland types and fish life histories are still rare. This study compares microplastic contamination in killifishes (Cyprinodontiformes: Rivulidae) with contrasting life histories—annual (short-lived: <italic>Notholebias minimus</italic> , <italic>Leptopanchax opalecens</italic> ) vs. perennial (long-lived: <italic>Kryptolebias ocellatus</italic> ; <italic>Kryptolebias hermaphroditus</italic> )—across two wetland types (temporary ponds vs. mangroves) on the coastal plain of Rio de Janeiro (Brazil). The tested hypothesis is that small, short-lived fishes in temporary wetlands exhibit lower microplastic contamination than perennial mangrove species, due to lower hydrological connectivity and shorter exposure time. Fishes were digested (KOH solution), vacuum filtered, and analysed using microscopy and µ-FTIR. Microplastic were detected in all species and 60.5% of individuals (1.58 ± 1.84 items fish⁻¹). Most particles were small (&lt; 1,000 µm), blue/black fragments or microfibers, with polymers dominated by polypropylene and poly(4-methyl-1-pentene). Contrary to H1, MP loads did not differ between mangroves and temporary ponds (GLMM: <italic>χ²</italic> = 0.18, <italic>p</italic> = 0.671), nor with body size ( <italic>χ²</italic> = 0.44, <italic>p</italic> = 0.507). Atmospheric deposition, precipitation, and runoff can supply rain-fed wetlands with MPs at levels sufficient to produce fish microplastic burdens comparable to those observed in tidally influenced mangroves. Convergent functional traits of rivulids—small gape, generalist foraging, and routine use of shallow microhabitats where fibers and fragments accumulate—likely equalize ingestion probabilities across life histories. Collectively, these findings show that temporary wetlands are not refuges from plastic contamination and should be explicitly included in monitoring and mitigation strategies that target diffuse, landscape-scale MP inputs.