A comparison of catchment travel times and storage deduced from deuterium and tritium tracers using StorAge Selection functions

Abstract

Abstract. Catchment travel time distributions (TTDs) are an efficient concept for summarizing the time-varying 3D transport of water and solutes towards an outlet in a single function of a water age and for estimating catchment storage by leveraging information contained in tracer data (e.g., deuterium 2H and tritium 3H). It is argued that the preferential use of the stable isotopes of O and H as tracers, compared to tritium, has truncated our vision of streamflow TTDs, meaning that the long tails of the distribution associated with old water tend to be neglected. However, the reasons for the truncation of the TTD tails are still obscured by methodological and data limitations. In this study, we went beyond these limitations and evaluated the differences between streamflow TTDs calculated using only deuterium (2H) or only tritium (3H). We also compared mobile catchment storage (derived from the TTDs) associated with each tracer. For this, we additionally constrained a model that successfully simulated high-frequency stream deuterium measurements with 24 stream tritium measurements over the same period (2015–2017). We used data from the forested headwater Weierbach catchment (42 ha) in Luxembourg. Time-varying streamflow TTDs were estimated by consistently using both tracers within a framework based on StorAge Selection (SAS) functions. We found similar TTDs and similar mobile storage between the 2H- and 3H-derived estimates, despite statistically significant differences for certain measures of TTDs and storage. The streamflow mean travel time was estimated at 2.90±0.54 years, using 2H, and 3.12±0.59 years, using 3H (mean ± 1 SD – standard deviation). Both tracers consistently suggested that less than 10 % of the stream water in the Weierbach catchment is older than 5 years. The travel time differences between the tracers were small compared to previous studies in other catchments, and contrary to prior expectations, we found that these differences were more pronounced for young water than for old water. The found differences could be explained by the calculation uncertainties and by a limited sampling frequency for tritium. We conclude that stable isotopes do not seem to systematically underestimate travel times or storage compared to tritium. Using both stable and radioactive isotopes of H as tracers reduced the travel time and storage calculation uncertainties. Tritium and stable isotopes both had the ability to reveal short travel times in streamflow. Using both tracers together better exploited the more specific information about longer travel times that 3H inherently contains due to its radioactive decay. The two tracers thus had different information contents overall. Tritium was slightly more informative than stable isotopes for travel time analysis, despite a lower number of tracer samples. In the future, it would be useful to similarly test the consistency of travel time estimates and the potential differences in travel time information contents between those tracers in catchments with other characteristics, or with a considerable fraction of stream water older than 5 years, since this could emphasize the role of the radioactive decay of tritium in discriminating younger water from older water.