Common streamflow dynamics unraveled the heavy-tailed flood distributions

Flood frequency distributions with heavy-tailed indicate a sizable chance of the occurrence of extreme floods. When heavy-tailed flood behavior is reliably identified, flood hazards caused by the unexpected can be reduced. However, for cases with limited or varying record lengths it is challenging to robustly estimate tail behavior with currently used indices, which rely solely on the graphical or mathematical performance of limited observations and are regardless of the physical processes.

In this work, we start by analyzing runoff generation processes and show that the hydrograph recession is a proper descriptor of the emergence of heavy-tailed behavior of flood frequency distributions. We then examine it in a large set of seasonal case studies, which encompasses a variety of climate and physiographic conditions across Germany. Our results show that the newly proposed approach can detect cases with heavy-tailed behavior, and compare severity across cases by evaluating their tail heaviness. Remarkably, it displays robust identification of heavy/nonheavy-tailed behavior for cases with short data records, benchmarked against two other frequently used metrics for heavy tails in hydrological studies, i.e., the upper tail ratio and the shape parameters of generalized extreme value distributions.

We highlight that the proposed method leverages the information of common discharge dynamics for inferring heavy-tailed flood behavior, which addresses the main limitations of currently used metrics and provides information on the characteristic flood hazard of river basins.

This study summarizes results of the DFG-funded project "Propensity of rivers to extreme floods: climate-landscape controls and early detection - PREDICTED" (Deutsche Forschungsgemeinschaft - German Research Foundation, Project Number 421396820).