Stream fishes have evolved ecological traits to adapt to various streamflow regimes. Humans are greatly modifying streamflow regimes through the construction of dams, land cover changes, water withdrawals, and changes in climate. Streamflow management to conserve fishes could be improved if we had a mechanistic understanding of how and why specific flows shape fish communities. An approach that specifies how a flow event will affect organisms with specific, mediating traits is an effective way to derive mechanistic understanding of flow-ecology relationships. Do to this, we use a dataset of more than 200 fish species distributions at stream reach resolution and a dataset of stream intermittency signatures at 540 stream gages to evaluate flow-ecology relationships in the conterminous United States. We hypothesize that locations with prolonged periods of no-flow will have (H1) decreased relative abundance of surface or water column feeders due to lower prey availability and (H2) decreased relative abundance of coldwater adapted species due to thermal stress. To analyze the flow-ecology relationships, we will standardize for community variation unrelated to hydrology (e.g. elevation, climate, and land cover) before developing models to relate fish ecology traits to low flow duration. By testing mechanistic hypotheses, we can derive predictable flow-ecology relationships that can be applied to other streams and management contexts.