In a hydrologically modified, warming and increasingly uncertain world, predicting and mitigating impacts of food web changes on ecosystem functioning and biodiversity is challenging. Understanding and accounting for time lags, legacies, extremes and shifting baselines is paramount. To this end, rivers that cyclically flood, cease flow and dry present multiple research opportunities as model ecosystems. Yet a review of the literature reveals such endeavours are rare. The majority (>65%) of food web research on intermittent rivers is conducted while they flow, with <25% of studies spanning multiple hydrological phases. Few consider antecedent conditions, from either lag or legacy perspectives. Less than 20% incorporate experimental methods necessary to provide hard evidence of causal relationships. There is also bias regarding the source of our understanding, with most studies conducted in dryland regions. Surprisingly, <30% of studies aim to investigate effects of flow intermittency on food webs, with <10% investigating drying effects specifically. Recent work shows that drying reduces aquatic food chain length and trophic diversity, often culminating in collapse of aquatic food webs, but also triggers switches in energy pathways and likely expands terrestrial food web dimensions. Global changes which exacerbate or produce unnatural drying may have deleterious and wide-ranging effects. Research on intermittent rivers, including the use of mesocosms and drying experiments, is key to develop and test hypotheses on and monitor and predict impacts of climate and hydrological changes on the food webs of dynamic ecosystems. We present several recommendations for how this can be achieved.