Lessons learned from the spatiotemporal analysis of long‐term and time‐variable young water fractions of large central European river basins

Abstract

AbstractThe transit time of precipitation entering a catchment and leaving it as streamflow spatiotemporally varies according to the flow paths that precipitation takes. However, investigating influences of hydrometeorological variables and catchment characteristics on time‐variable transit times is challenging due to the complex water flow through heterogeneous landscapes. Recent studies investigated the fraction of streamflow younger than approximately 3 months (Fyw) using multi‐year data (long‐term Fyw) or one‐year calculation windows to investigate its time‐variability (time‐variable Fyw). Nonetheless, it is still unclear if the inter‐annual variability of one‐year Fyw is due to hydrological influences or its uncertainty, and no minimum time series length for the long‐term Fyw was defined yet. Here, we investigated the impact of catchment characteristics and hydrometeorological variables on the long‐term Fyw, the time‐variable Fyw, and the Fyw depending on discharge of nine river basins in Central Europe. All methods of estimating Fyw led to similar results, with Fyw depending on discharge deemed unreliable due to the monthly sampling interval of streamflow isotopes. Danube and Rhine had the lowest Fyw (0.06), medium Fyw (0.20) were found in eastern basins (e.g., Oder), and the western ones (e.g., Mosel) had the highest Fyw (0.33). Spatial analysis indicated a negative relationship between Fyw and altitude. Contradicting or lacking spatiotemporal relationships to other variables pointed to unknown influential factors controlling the runoff process. Using flow duration curves, it was found that basins with a low flow variability had low Fyw, indicating that the old water of their runoff stems from large subsurface storage. With increasing calculation window size, the within‐basin variability of time‐variable Fyw decreased. Long‐term Fyw depended on the method used to define ‘long‐term’ and the time series length and start and end date. We thus recommend future studies to calculate long‐term and time‐variable Fyw to facilitate comparability between different catchments, and to account for this source of uncertainty. Further, more studies are needed in diverse catchments to investigate the hydrometeorological impacts on Fyw and thus on runoff generation processes.