Diverse yet Surprisingly Weak Influence of Snow-Fraction Changes on Regional Streamflow Seasonality Shifts

Snowmelt-driven streamflow is often highly seasonal and supports ecosystems and human water use. Climate-driven changes in snow accumulation and melt alter the timing and rates of liquid water input to catchments, thereby reshaping seasonal streamflow patterns. However, shifts in streamflow seasonality under climate change (e.g. snow fraction) remain uncertain. Here, we employ directional statistics to quantify streamflow seasonality (i.e., center of mass timing and concentration) and their sensitivity to annual snow fraction for 239 snow-affected CONUS catchments. While the snowfall-fraction sensitivity of streamflow timing and concentration is relatively weak in individual catchments, consistent and distinguishable patterns emerge at the regional scale. We demonstrate and explain an apparently opposite regional response of seasonal streamflow to between-year variations in snowfall fraction. In years with less snowfall, we identify regions of the USA where seasonal streamflow occurs later (Eastern Rockies and Great Plain-western Great Lakes) and where the flow becomes more concentrated in time (Pacific Northwest). These effects are precisely the opposite of the expected behaviour (observed in other snow-affected parts of the USA), which would be that less snowfall leads to earlier and less concentrated seasonal streamflow. The climate context, particularly precipitation seasonality, provides a mechanistic explanation for these unexpected behaviours. Further, trends from 1980 to 2022 show that changes in streamflow seasonality do not always match the expected effects of declining snow. Our results imply that climate change will not affect snow-affected water resources in the same way everywhere. Water managers in snow-affected regions will need to adapt their strategies to local climate conditions, taking into account not only changes in snow but also shifts in precipitation.