Host-associated microbiomes provide excellent means to explore eco-evolutionary processes that influence microbiome recruitment and dynamics of species compositions. Especially fish skin microbiomes emerge as a useful media by allowing experimental manipulations and noninvasive repeated sampling. In this study, we investigated the spatiotemporal dynamics of fish skin microbiomes by high throughput sequencing of amplicon sequence variants from metabarcoding of V3-V4 regions in the 16S rRNA gene. We used a wild-caught anadromous roach population (Rutilus rutilus) in the Baltic Sea as a model system to investigate the role of ecological filtering imposed by surrounding environmental factors, host characteristics, and host-microbe interactions. The experiment included several manipulations: i) we translocated individuals between fresh and brackish water, ii) treated with an antibacterial disinfectant to reboot microbial community, and iii) kept individuals alone or in pairs to explore the role of social environment and microbial dispersal between hosts. The analyses showed rapid shifts in microbiome composition as a result of ecological filtering imposed by translocations between different water environments. Host individual played a significant role in microbiome composition as inter-individual differences were repeatable in space and time whereas dispersal from water environment or nearby host did not play an important role in microbial recruitment and community composition.