Using isotopes to constrain water flux and age estimates in snow-influenced catchments using the STARR (Spatially distributed Tracer-Aided Rainfall–Runoff) model

Abstract

Abstract. Tracer-aided hydrological models are increasingly used to reveal fundamentals of runoff generation processes and water travel times in catchments. Modelling studies integrating stable water isotopes as tracers are mostly based in temperate and warm climates, leaving catchments with strong snow influences underrepresented in the literature. Such catchments are challenging, as the isotopic tracer signals in water entering the catchments as snowmelt are typically distorted from incoming precipitation due to fractionation processes in seasonal snowpack.We used the Spatially distributed Tracer-Aided RainfallRunoff (STARR) model to simulate fluxes, storage, and mixing of water and tracers, as well as estimating water ages in three long-term experimental catchments with varying degrees of snow influence and contrasting landscape characteristics. In the context of northern catchments the sites have exceptionally long and rich data sets of hydrometric data and -most importantly -stable water isotopes for both rain and snow conditions. To adapt the STARR model for sites with strong snow influence, we used a novel parsimonious calculation scheme that takes into account the isotopic fractionation through snow sublimation and snowmelt.The modified STARR setup simulated the streamflows, isotope ratios, and snow pack dynamics quite well in all three catchments. From this, our simulations indicated contrasting median water ages and water age distributions between catchments brought about mainly by differences in topography and soil characteristics. However, the variable degree of snow influence in catchments also had a major influence on the stream hydrograph, storage dynamics, and water age distributions, which was captured by the model. Our study suggested that snow sublimation fractionation processes can be important to include in tracer-aided modelling for catchments with seasonal snowpack, while the influence of fractionation during snowmelt could not be unequivocally shown. Our work showed the utility of isotopes to provide a proof of concept for our modelling framework in snowinfluenced catchments.

Water ages in the critical zone of long-term experimental sites in northern latitudes

“…Catchment runoff ages and plot scale AR show also similar dynamics at Dorset, as Piovano et al (in review) also reported generally increasing water ages in the runoff during the snow dominated winter, while runoff ages during summer were more dependent on the rainfall pattern. We see also similarities at Krycklan with increasing water ages for catchment runoff 15 (Ala-aho et al, 2017b) and soil water recharge throughout the snow dominated winter. However, while snowmelt leads to a rapid decrease in water ages at both the catchment and the plot scales, the cessation of the soil recharge flux in the summer in the 1D-simulations results in the largest AR during fall when older soil water becomes remobilized.…”

“…The R water ages at 50 cm depth in this study are obviously younger than catchment scale runoff water ages Ala-aho et al, 2017b;Benettin et al, 2017;Kuppel et al, 2018). Nevertheless, there are similarities regarding the age dynamics.…”

“…Nevertheless, there are similarities regarding the age dynamics. The plot scale R water ages based on SWIS show a similar pattern to catchment runoff 10 water ages with generally increasing values throughout spring towards summer (decreasing storage) and lowest water ages during winter (highest storage) Ala-aho et al, 2017b;Benettin et al, 2017). Catchment runoff ages and plot scale AR show also similar dynamics at Dorset, as Piovano et al (in review) also reported generally increasing water ages in the runoff during the snow dominated winter, while runoff ages during summer were more dependent on the rainfall pattern.…”

“…In contrast, Ala-aho et al (2017b) reported generally small runoff water ages during summer at Krycklan with few peaks due to dry periods, probably due to young water contributions from small mires (Peralta-Tapia et al, 2015). 20…”

“…Our plot scale soil hydraulic simulations can help to better understand the processes taking place within the catchment and constraining or benchmarking spatially distributed hydrological models (van Huijgevoort et al, 2016;Ala-aho et al, 2017b;Kuppel et al, 2018). Such catchment models cannot account for the soil physical processes in a similar detail 25 as a 1D-model due to computational limitations.…”