River regulation has resulted in substantial losses in habitat connectivity, biodiversity and ecosystem services. River managers are faced with a growing need to protect the key aspects of the natural flow regime. A practical approach to providing environmental flow standards is to create a regional framework by classifying unregulated streams into groups of similar hydrologic properties, which represent natural flow regime targets. Because spatial resolution can influence the structure of regional datasets, it may be advantageous to relate datasets created at different scales in order to establish hierarchical structure and to understand how the relative importance of variables change with regard to scale. The purpose of this study was to classify unregulated streams within an eight-state region into groups in order to provide environmental flow standards for managers and to relate that dataset to frameworks created at larger scales. Using USGS daily stream gauge information, we used 66 hydrologic statistics to classify 292 streams in groups of similar hydrologic properties. We isolated six flow classes in a sub-region of the Southeastern US that ranged from extremely stable to highly variable to intermittent. We developed classification trees to reduce the number of hydrologic variables for future classifications. By comparing flow classes in our study to those of the entire US, we found that hierarchical structure did exist and that the divergence of flow classes will largely depend on the spatial resolution. Copyright © 2011 John Wiley & Sons, Ltd.