Validation of SNOWPACK Modelled Densities Using SnowMicroPen Data

Snow density is a key parameter influencing snowpack stability, hydrology, and energy balance. Accurate simulation of its evolution is critical for applications ranging from avalanche forecasting to climate modelling. This thesis evaluates the performance of the SNOWPACK model in predicting snow density by comparing its outputs to the high resolution density profiles derived from SnowMicroPen (SMP) measurements collected daily over nine winters (2015–2024) at the Weissfluhjoch field site. The SMP data were first validated using reference methods including density cutters and manual snow water equivalent measurements. A refined parametrization to obtain density from SMP force profile, was developed and tested, considering sensor offsets and long term bias. Modelled new snow densities from SNOWPACK were then assessed against SMP-derived densities at the time of deposition across over 1800 snowfall layers. Results show that SNOWPACK tends to underestimate newsnowdensity under warm conditions and overestimate under colder, windier events, partially due to discontinuities in its current parametrization. A new empirical parametrization based on SMP data and meteorological inputs was proposed to address this, improving RMSE by 10 kg/m3 and reducing seasonal biases, particularly in spring. However, the range of predicted new snow densities was unrealistically narrow, failing to reproduce the full variability observed in field measurements. Additionally, the seasonal evolution of eight snowlayers across twowinters was tracked in both SMP and SNOWPACK data. While general settling trends were consistent, deviations were event-specific and highlighted limitations in layer matching and temporal resolution. Overall, this study underscores the value of highresolution SMP data for validating and improving snow density models and proposes tracks to enhance SNOWPACK’s performance.