Urbanization impacts land, air, and water, creating environmental gradients between cities and rural areas. Urban stormwater delivers myriad co-occurring, understudied, and mostly unregulated contaminants to aquatic ecosystems, causing a pollution gradient. Recipient ecosystems host interacting species that can affect each others’ growth and responses to these contaminants. For example, plants and their microbiomes often reciprocally increase growth and contaminant tolerance. Here, we identified ecological variables affecting contaminant fate across an urban-rural gradient using 50 sources of the aquatic plant Lemna minor (duckweed) and associated microbes, and two co-occurring winter contaminants of temperate cities, benzotriazole and salt. We conducted experiments totalling 2,500 independent host-microbe-contaminant microcosms. Benzotriazole and salt negatively affected duckweed growth, but not microbial growth, and duckweeds maintained faster growth with their local, rather than disrupted, microbiota. Benzotriazole transformation products of plant, microbial, and phototransformation pathways were linked to duckweed and microbial growth, and were affected by salt co-contamination, microbiome disruption, and source sites of duckweeds and microbes. Duckweeds from urban sites grew faster and enhanced phytotransformation, but supported less total transformation of benzotriazole. Increasing microbial community diversity correlated with greater removal of benzotriazole, but taxonomic groups may explain shifts across transformation pathways: the genus Aeromonas was linked to increasing phototransformation. Because benzotriazole toxicity could depend on amount and type of in situ transformation, this variation across duckweeds and microbes could be harnessed for better management of urban stormwater. Broadly, our results demonstrate that plant-microbiome interactions harbour manipulable variation for bioremediation applications.