Comparing Satellite-Derived and Isotope-Based Estimates of Snow Contribution to Runoff in an Alpine Catchment

Despite the importance of snow contributions to the water resources collected by alpine catchments, their precise measurement and monitoring remain challenging due to their complexity and inaccessibility. Stable Isotopes of water (δ¹⁸O, δ²H, and δ¹⁷O) allow separation of streamflow into water that entered the catchment as snow versus rain. Meanwhile, progress in process-based simulations of snow (physics-based FSM2oshd model) fused with satellite snow cover data has enhanced the accuracy of gridded data products such as snow water equivalent (SWE) and runoff from snow melt (ROS) in mm at a resolution of 250m (Mott, 2023).

Between June 9, 2016, and September 24, 2018, 2548 water samples from the Avançon de Nant (Western Swiss Alps, 13.4 km², 1200 to 3051 m a.s.l.; see Michelon et al., 2023) were analyzed for δ¹⁸O, δ²H, and δ¹⁷O and compared with 157 snow and 95 rain samples taken in the catchment during the same period. A simple mixing model of snow and rain was used to estimate the porporation of daily discharge originating from snow (Psnow). Over the same period, the discharge (Q) was measured and and multiplied with Psnow to estimate discharge from snow, Qsnow.

There is a clear seasonality of the correspondence between the ROS and Qsnow: during the low flow period, Qsnow exceeds ROS. In contrast, during the peak flow periods, e.g., during the spring “freshet” period, ROS exceeds Qsnow. Their correlation is statistically significant during the spring freshet (April–June), because direct snow runoff, hving undergone minimal storage, dominates the streamflow. When snow-free periods are excluded, the Qsnow, as determined by isotopes, is more correlated with the ROS than is the total Q. This difference is obscured when including ROS-free periods. The discrepancy we see can be explained by the fact that ROS does not account for storage and release beyond the grid scale, namely the catchment scale, and thus may eventually be the basis for a travel time calculation.

This example in a single catchment allows inter-scale comparisons, moving beyond validation to developing larger-scale monitoring tools. Collecting and analyzing stable isotope samples is labor-intensive and not universally possible. Thus, finding tools that enable the information they deliver to be gleaned from other sources is very useful. Ongoing work compares how these comparisons vary according to other satellite-derived products, such as SWE, at a higher resolution, which may be more ubiquitous. 

References

Michelon, A., Ceperley, N., Beria, H., Larsen, J., Vennemann, T., and Schaefli, B.: Hydrodynamics of a high Alpine catchment characterized by four natural tracers, Hydrol. Earth Syst. Sci., 27, 1403–1430, https://doi.org/10.5194/hess-27-1403-2023, 2023.

Mott, R.: Seasonal snow data for Switzerland OSHD - FSM2sohd (1.0), https://doi.org/10.16904/ENVIDAT.404, 2023.