Verification of the multi-layer SNOWPACK model with different water transport schemes
- Journal Article
Rights / licenseCreative Commons Attribution 3.0 Unported
The widely used detailed SNOWPACK model hasundergone constant development over the years. A notablerecent extension is the introduction of a Richards equation(RE) solver as an alternative for the bucket-type approachfor describing water transport in the snow and soil layers.In addition, continuous updates of snow settling and newsnow density parameterizations have changed model behav-ior. This study presents a detailed evaluation of model per-formance against a comprehensive multiyear data set fromWeissfluhjoch near Davos, Switzerland. The data set is col-lected by automatic meteorological and snowpack measure-ments and manual snow profiles. During the main win-ter season, snow height (RMSE:<4.2 cm), snow waterequivalent (SWE, RMSE:<40 mm w.e.), snow tempera-ture distributions (typical deviation with measurements:<1.0◦C) and snow density (typical deviation with observa-tions:<50 kg m−3) as well as their temporal evolution arewell simulated in the model and the influence of the two wa-ter transport schemes is small. The RE approach reproducesinternal differences over capillary barriers but fails to predictenough grain growth since the growth routines have been cal-ibrated using the bucket scheme in the original SNOWPACKmodel. However, the agreement in both density and grainsize is sufficient to parameterize the hydraulic properties suc-cessfully. In the melt season, a pronounced underestimationof typically 200 mm w.e. in SWE is found. The discrepan-cies between the simulations and the field data are generallylarger than the differences between the two water transport schemes. Nevertheless, the detailed comparison of the inter-nal snowpack structure shows that the timing of internal tem-perature and water dynamics is adequately and better repre-sented with the new RE approach when compared to the con-ventional bucket scheme. On the contrary, the progress of themeltwater front in the snowpack as detected by radar and thetemporal evolution of the vertical distribution of melt formsin manually observed snow profiles do not support this con-clusion. This discrepancy suggests that the implementationof RE partly mimics preferential flow effects. Show more
Journal / seriesThe Cryosphere
Pages / Article No.
Organisational unit03953 - Robertsson, Johan / Robertsson, Johan
MoreShow all metadata