The climate change often affects the variability and persistence of river discharges that may show an alterated balance between snow and rainfall and an intensification of extreme hydrological events. Such climate-induced hydrologic changes may have relevant consequences on the freshwater ecosystem. For this reason, the search of simple but effective tools for river regime classification is still a topic of interest in order to investigate variations in flow regimes, and evaluate future climate impact.

In this article, we present a procedure for classifying streamflow time series according to their underlying dynamic structures and we show its effectiveness analyzing observed streamflow data of 221 catchments in the United States.