A multi-step approach for streamflow classification

The article presents a strategy for classifying streamflows into groups based on their temporal dynamics. It compares dynamic patterns of river discharges over time, capturing seasonality and short-term components using linear time series models. The final classification offers a means of identifying hydrological regimes and gives a preliminary understanding of the potential impacts of climate change on streamflows. The approach is illustrated by analyzing streamflow data from 221 stations in the United States in two non overlapping periods.