Advancing isotope-enabled snowpack modelling: Development and evaluation at a boreal-subarctic site

Stable water isotopes (i.e., δ¹⁸O, δ²H) are key tracers of water sources, flow pathways, and transport processes in hydrological systems. In snow-dominated regions, seasonal snowpacks strongly control water storage, release, and snowmelt isotope signals used in tracer-aided analyses. However, the isotopic evolution of seasonal snowpacks is still poorly represented in hydrological models. Most previous models therefore assume that snowmelt isotopic composition equals that of the snowpack, despite observations showing systematic differences during the melt season. These differences can not be resolved without explicitly representing isotope fractionation and are further constrained by the limited availability of high-resolution isotope data. The objective of this study is therefore to improve the simulation of snowpack and snowmelt isotope dynamics by developing and systematically evaluating a physically based, isotope-enabled multi-layer snowpack model (FSM-Iso) for a boreal subarctic environment. The analysis is based on high-temporal-resolution observations of snowpack and snowmelt isotopic composition, snow water equivalent, and meteorological forcing from northern Finland (Pallas site). FSM-Iso explicitly couples stable water isotope evolution to snow, isotope mass and energy balance and represents isotope fractionation during sublimation, melting, and refreezing using physically based formulations. Model performance is evaluated through a stepwise comparison with two widely used isotope-enabled snowpack models that span a range from conceptual to simplified physically based approaches. Results show that FSM-Iso reproduces observed seasonal isotope dynamics in both the snowpack and snowmelt well, including the transition from isotopically depleted early meltwater to progressively enriched meltwater later in the melt season. In contrast, simplified snow isotope models systematically misrepresent both the timing and magnitude of meltwater isotope enrichment, resulting in biased snowmelt isotope signals. These results demonstrate that coupling isotope fractionation processes consistently with mass and energy balance is critical for reliable tracer-aided hydrological modelling in snow-dominated catchments.