Scaling Precipitation Input to Spatially Distributed Hydrological Models by Measured Snow Distribution

Accurate knowledge on snow distribution in alpine terrain is crucial for various applications such as flood risk assessment, avalanche warning or managing water supply and hydro-power. To simulate the seasonal snow cover development in alpine terrain, the spatially distributed, physics-based model Alpine3D is suitable. The model is typically driven by spatial interpolations of observations from automatic weather stations (AWS), leading to errors in the spatial distribution of atmospheric forcing. With recent advances in remote sensing techniques, maps of snow depth can be acquired with high spatial resolution and accuracy. In this work, maps of the snow depth distribution, calculated from summer and winter digital surface models based on Airborne Digital Sensors (ADS), are used to scale precipitation input data, with the aim to improve the accuracy of simulation of the spatial distribution of snow with Alpine3D. A simple method to scale and redistribute precipitation is presented and the performance is analyzed. The scaling method is only applied if it is snowing. For rainfall the precipitation is distributed by interpolation, with a simple air temperature threshold used for the determination of the precipitation phase. It was found that the accuracy of spatial snow distribution could be improved significantly for the simulated domain. The standard deviation of absolute snow depth error is reduced up to a factor 3.4 to less than 20 cm. The mean absolute error in snow distribution was reduced when using representative input sources for the simulation domain. For inter-annual scaling, the model performance could also be improved, even when using a remote sensing dataset from a different winter. In conclusion, using remote sensing data to process precipitation input, complex processes such as preferential snow deposition and snow relocation due to wind or avalanches, can be substituted and modeling performance of spatial snow distribution is improved.

Modelling Drifting Snow Sublimation and Investigating Small-Scale Snow Transport

Christine Dorothea Groot Zwaaftink

Drifting snow has a large influence on the mountain snow cover and the overlying atmospheric boundary layer. Consequently, drifting snow also affects avalanches, snow hydrology, vegetation and climate. Considerable effort is devoted to understanding the process in order to model the effects on the mountain snow cover or the overlying atmosphere. This thesis work aims to I) provide estimates of drifting snow sublimation within a small catchment in the Swiss Alps and II) test an alternative method to describe snow transport on short time and small spatial scales. Drifting snow sublimation moistens and cools the surrounding air and causes a loss of snow mass. Based on the feedback mechanisms integrating air temperature, humidity and snow concentration, it should be seen as a self-limiting process. Drifting snow sublimation (including these feedback mechanisms) has been implemented in Alpine3D, which is a high-resolution snow transport model for alpine terrain. This model has been applied to a small area in the Swiss Alps for a case study of approximately two days. Results show that, in general, the negative feedbacks of sublimation on the snow mass concentration, air temperature, and humidity are small. They are, however, relevant on the slope scale, as could be retrieved from analyzing the deposition on a leeward slope. Compared to a reference simulation without sublimation, drifting snow sublimation reduced the deposition by approximately 12% on this leeward slope when omitting all feedback mechanisms. In a simulation where all feedbacks were included, drifting snow sublimation reduced the snow deposition on this slope by 10%. Drifting snow sublimation thus appears to be a significant process for a leeward slope influenced by drifting snow. On the contrary, the spatially averaged reduction of deposition due to drifting snow sublimation in the complete modelled domain was 2.3%. During this case study, this value was comparable to surface sublimation. At the same field site a simulation of a complete season was performed and in this case, sublimation of drifting snow was accounted for only in the absence of concurrent precipitation. As for the case study, the slope affected the most by drifting snow sublimation was shown to be a leeward slope for SE wind. This could be explained by generally warmer and dryer conditions during periods with SE wind. In this particular slope, the snow amount was reduced by about 1.8% due to drifting snow sublimation. For the complete domain however, the model results indicate that drifting snow sublimation is much smaller than surface sublimation and negligible for the total water equivalent. Thus, at the studied field site drifting snow sublimation seems only significant locally or on short time scales. Limitations of drifting snow simulations based on mean wind fields are seen as the variability of the snow cover is underestimated on scales of few meters. Moreover, these models are also not able to represent snow transport on very short time scales. High-frequency observations of drifting snow and sand saltation have shown that turbulent structures have an influence on the snow or sand mass flux. Improvements in modelling drifting snow on such time scales can thus be reached by considering turbulence. Here, a combination of wind fields from large eddy simulations and a Lagrangian stochastic model was used to describe snow particle trajectories over flat terrain. This model reproduced a time series of the snow mass flux that qualitatively corresponds to observations from the Weissfluhjoch Versuchsfeld. Furthermore, results show that the use of the surface shear stress rather than the friction velocity, as in conventional models, resulted in a spatial variation of drifting snow despite homogeneous snow properties and the flatness of the considered terrain. The distribution of the surface shear stress in space and time was found reasonable yet narrow based on the ratio between the standard deviation and the mean shear stress, which was 0.17 compared to a value of 0.4 reported for measurements. Nonetheless, this variable surface shear stress played a key role in the modelled spatial variation of drifting snow. This also points at a relevant influence of the particle aerodynamic entrainment process for drifting snow. The results thus show that this model seems to overcome limitations of conventional three-dimensional snow transport models as it gives a more realistic description of drifting snow on short time scales and it may be used to gain further insight in drifting snow.

EPFL2013

Spatial variability in snow precipitation and accumulation in COSMO–WRF simulations and radar estimations over complex terrain

Alexis Berne, Nikola Besic, Marco Gabella, Franziska Gerber, Michael Lehning, Varun Sharma

Snow distribution in complex alpine terrain and its evolution in the future climate is important in a variety of applications including hydropower, avalanche forecasting and freshwater resources. However, it is still challenging to quantitatively forecast precipitation, especially over complex terrain where the interaction between local wind and precipitation fields strongly affects snow distribution at the mountain ridge scale. Therefore, it is essential to retrieve high-resolution information about precipitation processes over complex terrain. Here, we present very-high-resolution Weather Research and Forecasting model (WRF) simulations (COSMO-WRF), which are initialized by 2.2 km resolution Consortium for Small-scale Modeling (COSMO) analysis. To assess the ability of COSMO-WRF to represent spatial snow precipitation patterns, they are validated against operational weather radar measurements. Estimated COSMO-WRF precipitation is generally higher than estimated radar precipitation, most likely due to an overestimation of oro-graphic precipitation enhancement in the model. The high precipitation amounts also lead to a higher spatial variability in the model compared to radar estimates. Overall, an autocorrelation and scale analysis of radar and COSMO-WRF precipitation patterns at a horizontal grid spacing of 450 m show that COSMO-WRF captures the spatial variability normalized by the domain-wide variability in precipitation patterns down to the scale of a few kilometers. However, simulated precipitation patterns systematically show a lower variability on the smallest scales of a few hundred meters compared to radar estimates. A comparison of spatial variability for different model resolutions gives evidence for an improved representation of local precipitation processes at a horizontal resolution of 50 m compared to 450 m. Additionally, differences of precipitation between 2830 m above sea level and the ground indicate that near-surface processes are active in the model.

2018

Evaluating a prediction system for snow management

Michael Lehning

The evaluation of snowpack models capable of accounting for snow management in ski resorts is a major step towards acceptance of such models in supporting the daily decision-making process of snow production managers. In the framework of the EU Horizon 2020 (H2020) project PROSNOW, a service to enable real-time optimization of grooming and snow-making in ski resorts was developed. We applied snow management strategies integrated in the snowpack simulations of AMUNDSEN, Crocus, and SNOWPACK-Alpine3D for nine PROSNOW ski resorts located in the European Alps. We assessed the performance of the snow simulations for five winter seasons (2015-2020) using both ground-based data (GNSS-measured snow depth) and spaceborne snow maps (Copernicus Sentinel-2). Particular attention has been devoted to characterizing the spatial performance of the simulated piste snow management at a resolution of 10 m. The simulated results showed a high overall accuracy of more than 80% for snow-covered areas compared to the Sentinel-2 data. Moreover, the correlation to the ground observation data was high. Potential sources for local differences in the snow depth between the simulations and the measurements are mainly the impact of snow redistribution by skiers; compensation of uneven terrain when grooming; or spontaneous local adaptions of the snow management, which were not reflected in the simulations. Subdividing each individual ski resort into differently sized ski resort reference units (SRUs) based on topography showed a slight decrease in mean deviation. Although this work shows plausible and robust results on the ski slope scale by all three snowpack models, the accuracy of the results is mainly dependent on the detailed representation of the real-world snow management practices in the models. As snow management assessment and prediction systems get integrated into the workflow of resort managers, the formulation of snow management can be refined in the future.

COPERNICUS GESELLSCHAFT MBH2021