Demonstrating and predicting the effectiveness of environmental watering to maintain ecosystem health is becoming increasingly important, notably in semi-arid floodplain ecosystems. Modelling plant floodplain vegetation responses to inform environmental watering outcomes is also challenging. We evaluated the effects of a large-scale environmental flow event on a semi-arid floodplain lakeside plant assemblage for meeting the management goal of increasing water-dependant taxa and functional groups. We developed a multi-taxon Bayesian hierarchical model to describe temporal and small-scale spatial patterns in taxonomic occurrences. We then examined community summary metrics (proportion of native taxa, plant functional groups) to evaluate patterns for the entire floodplain lakes system, the scale most relevant to management. Overall, in a system dominated by terrestrial dry plant taxa, 52.9% of terrestrial damp plant taxa showed a short-term increase in occurrence in response to the environmental flow, which translated into similar responses in some functional groups. However, nearly half of the plant taxa that increased then demonstrated a decline by 18 months after the flow event. Our community summary metrics captured these general results; however, they were disproportionately influenced by a few abundant plant taxa. These results highlight the advantages of multi-taxon models for interpreting flow responses and developing effective environmental flow management strategies, because they can be used to summarise community responses, while preserving important taxon-specific information. In semi-arid systems, where river regulation and climate change have reduced the frequency of flood events, the ability to deliver environmental flows during protracted periods of drought may be a policy option to restore or maintain the natural floodplain vegetation assemblage and prevents the transition to dryland taxa. © 2018 John Wiley & Sons Ltd.