Can we quantify the fraction of temporary streams in river networks?

The temporary fraction, denoted as Tf , represents the ratio between the length of temporary streams and the total length of a river network. Estimating regional-scale temporary fractions in large, branching networks is challenging due to the scarcity of empirical data and the tendency of modeling approaches to overlook headwater streams. However, understanding how the temporary fraction varies across different scales and determining the proportion of large-scale river systems composed of temporary channels is crucial. This knowledge has significant implications for assessing river vulnerability to climate change and developing targeted conservation and management strategies for intermittent rivers. In this study, we integrate experimental data, theoretical analysis, and modeling results to demonstrate that the temporary fraction exhibits significant scaling with the contributing area, meaning it depends on the resolution used to define the network domain. In particular, empirical data collected in 10 headwaters catchments located in Italy and in the U.S.A., and large-scale modelling results suggest that the headwaters are more prone to flow intermittency as compared to larger channels. Our analysis also indicates that: i) high-order channels bear a poor informative content for the estimate of the total fraction of temporary streams in a river, as that the headwaters are comparatively longer and more dynamic; ii) temporary fractions in different catchments or basins might be much higher than previously estimated: for the Italian catchments Tf ranges from 0.56 to 0.95, with a regional-scale prediction of 0.6 − 0.8, while the US basins’ Tf ranges from 0.23 to 0.87, with a regional-scale prediction that approaches 0.8. This work offers a new framework for estimating the percentage of temporary channels within different types of river basins and for extrapolating empirical data or modelling results across scales.