In their effort to balance anthropogenic water demand and ecosystem conservation within a sustainable water resources management framework, water managers and stakeholders need sound scientific guidance. In this study, we applied a two-dimensional hydrodynamic habitat model using benthic macroinvertebrates as the target aquatic community and carried out an environmental flow (eflow) assessment downstream of the Marathon Reservoir (Attica, Central Greece). Hydrology-based eflow scenarios were additionally developed over an 11-year period, and the lowest acceptable ecosystem-based eflow was compared with the hydrology-based eflow predictions. We found that the hydrological methods tend to underestimate the eflows required to ensure functional aquatic ecosystems. The results showed that (a) the different hydrological methods developed highly variable eflow scenarios, ranging from 0.0006 to 0.18 m 3 /s, (b) the ecosystem-based eflow was up to 183% higher than the hydrology-based ones and 26% to 465% higher than those defined by the national legislation, and (c) the probability of agreement between hydrological and ecological predictions was 12.5%, as only one out of the eight hydrology-based scenarios coincided with the ecosystem-based eflows. We conclude that hydrological methods should be used with caution in the absence of ecological information. Their use as stand-alone tools seems problematic and bears a high risk of producing inappropriate eflow scenarios. Integrative frameworks combining hydrological-ecological methods could be useful to provide information on what is ecologically acceptable and hydrologically socially feasible, but because the two methods comprise structurally different, noninteracting concepts, they are inherently insufficient to increase the confidence of predicting and selecting eflows.