Dr. Charles Fierz

From the start, the SNOWPACK model was created to address the needs of the avalanche warning in Switzerland based on a large network of Automatic Weather Stations (AWS). An energy balance approach was chosen that incorporated the modeling of the snow metamorphism in order to reproduce the stratigraphy, an arbitrary number of layers and the ability to simulate very thin layers. But a number of annoying issues remained for the daily use in an operational toolchain, linked to the handling of raw data coming from the AWS as well as the limitations of the AWS themselves. After the data preprocessing got separated from the SNOWPACK model itself as the MeteoIO library (in 2008), it experienced a sustained development effort over more than 15 years. Many typical data issues could be identified and addressed, often with timeseries filters in MeteoIO while a few had to be addressed in the SNOWPACK model itself. More flexibility was also allowed by new developments aimed at supporting operational constraints such as the inability to physically access some of the AWS in the winter, even to perform emergency maintenance.
This paper shows some of the limitations of the measurement network from the SNOWPACK point of view and some typical data problems that have been encountered in the last 15 years while forcing SNOWPACK with raw AWS data. Some of the solutions that have been implemented and the compromises that had to be made are presented as well as their remaining limitations.

Water percolation in snow plays a crucial role in the avalanche risk assessment. Liquid water content and wetting front are hard to measure in the field; hence, accurate simulation of the phenomena can be of great help to forecasters. This study was the first to evaluate water percolation simulations with the SNOWPACK model using Richards’ scheme on Mount Fidelity, Glacier National Park, Canada. The study highlights that, at this site, an updated configuration on precipitation phase transition and new snow density can significantly improve simulations of the snow cover, and water percolation in particular, which can be relevant in an era of an increased occurrence of rain-on-snow (ROS) events. More specifically, emphasis was put on the quality of the input data and parameters. The analysis of the precipitation phase temperature threshold showed that a value of 1.4°C was the best suited to track the rain/snow transition on site. A 10-year analysis of 24-h precipitation measured using the rain gauge and 24-h new snow water equivalent showed an excellent correlation. New snow density sub-models were evaluated using the 24-h new snow density values taken by the park technicians. The BELLAIRE model performed best and was used to drive the snow simulations. Two SNOWPACK snow simulations were evaluated using 1) rain gauge precipitation amount (PCPM) and 2) automatic snow height measurement (HS) at the same site. Both runs simulated the main snowpack layers observed during the dry season (i.e., before spring percolation was observed), and both simulated the snow properties with good accuracy. The water equivalent of snow cover, used as a proxy for a first-order characterization of the simulations generated by both simulations, was slightly underestimated compared with four manual measurements taken on-site during the winter. Nevertheless, the comparison of both measured density and modeled bulk density showed great correspondence. The percolation timing and wetting front depth were evaluated using field measurements from field campaigns and continuous observations from on-site instruments. The main percolation events were correctly simulated and were coincident with the observed wet avalanche cycles. The results highlight the need for accurate input data on valid simulation of the wetting front and percolation timing on site. Good percolation information generated using the SNOWPACK model and Richards’ scheme could be used to assess the snowpack stability by forecasters in areas where such data are available.

The necessity of characterizing snow through objective, physically motivated parameters has led to new model formulations and new measurement techniques. Consequently, essential structural parameters such as density and specific surface area (for basic characterization) or mechanical parameters such as the critical crack length (for avalanche stability characterization) gradually replace the semiempirical indices acquired from traditional stratigraphy. These advances come along with new demands and potentials for validation. To this end, we conducted the RHOSSA field campaign, in reference to density (ρ) and specific surface area (SSA), at the Weissfluhjoch research site in the Swiss Alps to provide a multi-instrument, multi-resolution dataset of density, SSA and critical crack length over the complete winter season of 2015-2016. In this paper, we present the design of the campaign and a basic analysis of the measurements alongside predictions from the model SNOWPACK. To bridge between traditional and new methods, the campaign comprises traditional profiles, density cutter, IceCube, SnowMicroPen (SMP), micro-computed-tomography, propagation saw tests and compression tests. To bridge between different temporal resolutions, the traditional weekly to biweekly (every 2 weeks, used in this sense throughout the paper) snow pits were complemented by daily SMP measurements. From the latter, we derived a recalibration of the statistical retrieval of density and SSA for SMP version 4 that yields an unprecedented spatiotemporal picture of the seasonal evolution of density and SSA in a snowpack. Finally, we provide an intercomparison of measured and modeled estimates of density and SSA for four characteristic layers over the entire season to demonstrate the potential of high-temporal-resolution monitoring for snowpack model validation.

Manually collected snow data are often considered as ground truth for many applications such as climatological or hydrological studies. However, there are many sources of uncertainty that are not quantified in detail. For the determination of water equivalent of snow cover (SWE), different snow core samplers and scales are used, but they are all based on the same measurement principle. We conducted two field campaigns with 9 samplers commonly used in observational measurements and research in Europe and northern America to better quantify uncertainties when measuring depth, density and SWE with core samplers. During the first campaign, as a first approach to distinguish snow variability measured at the plot and at the point scale, repeated measurements were taken along two 20 m long snow pits. The results revealed a much higher variability of SWE at the plot scale (resulting from both natural variability and instrumental bias) compared to repeated measurements at the same spot (resulting mostly from error induced by observers or very small scale variability of snow depth). The exceptionally homogeneous snowpack found in the second campaign permitted to almost neglect the natural variability of the snowpack properties and focus on the separation between instrumental bias and error induced by observers. Reported uncertainties refer to a shallow, homogeneous tundra‐taiga snowpack less than 1 meter deep (loose, mostly recrystallized snow and no wind impact). Under such measurement conditions, the uncertainty in bulk snow density estimation is about 5% for an individual instrument and is close to 10% among different instruments. Results confirmed that instrumental bias exceeded both the natural variability and the error induced by observers, even in the case when observers were not familiar with a given snow core sampler.

Twenty-seven models participated in the Earth System Model - Snow Model Intercomparison Project (ESM-SnowMIP), the most data-rich MIP dedicated to snow modelling. Our findings do not support the hypothesis advanced by previous snow MIPs: evaluating models against more variables, and providing evaluation datasets extended temporally and spatially does not facilitate identification of key new processes requiring improvement to model snow mass and energy budgets, even at point scales. In fact, the same modelling issues identified by previous snow MIPs arose: albedo is a major source of uncertainty, surface exchange parametrizations are problematic and individual model performance is inconsistent. This lack of progress is attributed partly to the large number of human errors that led to anomalous model behaviour and to numerous resubmissions. It is unclear how widespread such errors are in our field and others; dedicated time and resources will be needed to tackle this issue to prevent highly sophisticated models and their research outputs from being vulnerable because of avoidable human mistakes. The design of and the data available to successive snow MIPs were also questioned. Evaluation of models against bulk snow properties was found to be sufficient for some but inappropriate for more complex snow models whose skills at simulating internal snow properties remained untested. Discussions between the authors of this paper on the purpose of MIPs revealed varied, and sometimes contradictory, motivations behind their participation. These findings started a collaborative effort to adapt future snow MIPs to respond to the diverse needs of the community.

The application of numerical modelling of the snowpack in support of avalanche hazard prediction is increasing. Modelling, in complement to direct observations and weather forecasting, provides information otherwise unavailable on the present and future state of the snowpack and its mechanical stability. However, there is often a perceived mismatch between the capabilities of modelling tools developed by research organizations and implemented by some operational services, and the actual operational use of those by avalanche forecasters. This causes frustration on both sides. By summarizing currently implemented modelling tools specifically designed for avalanche forecasting, we intend to diminish and contribute to bridging this gap. We highlight specific features and potential added value, as well as challenges preventing a more widespread use of these modelling tools. Lessons learned from currently used methods are explored and provided, as well as prospects for the future, including a list of the most critical issues to be addressed.

Ice layers may form deep in the snowpack due to preferential water flow, with impacts on the snowpack mechanical, hydrological and thermodynamical properties. This detailed study at a high-altitude alpine site aims to monitor their formation and evolution thanks to the combined use of a comprehensive observation dataset at a daily frequency and state-of-the-art snow-cover modeling with improved ice formation representation. In particular, daily SnowMicroPen penetration resistance profiles enabled us to better identify ice layer temporal and spatial heterogeneity when associated with traditional snowpack profiles and measurements, while upward-looking ground penetrating radar measurements enabled us to detect the water front and better describe the snowpack wetting when associated with lysimeter runoff measurements. A new ice reservoir was implemented in the one-dimensional SNOWPACK model, which enabled us to successfully represent the formation of some ice layers when using Richards equation and preferential flow domain parameterization during winter 2017. The simulation of unobserved melt-freeze crusts was also reduced. These improved results were confirmed over 17 winters. Detailed snowpack simulations with snow microstructure representation associated with a high-resolution comprehensive observation dataset were shown to be relevant for studying and modeling such complex phenomena despite limitations inherent to one-dimensional modeling.

The International Association of Cryospheric Sciences (IACS) became the eighth and most recent association of IUGG at the general assembly in Perugia, Italy, in July 2007. IACS was launched in recognition of the importance of the cryosphere within the Earth system, particularly at a time of significant global change. It was the first new association of the union to be formed in over 80 years and IACS celebrated its 10th anniversary only a year before the IUGG centennial. The forbearers of IACS, however, stretch back even further than IUGG, starting with the formation of the Commission Internationale des Glaciers (CIG) by the International Geological Congress in 1894. Here we record the history of the transition from CIG to IACS, the scientific objectives that drove activities and changes, and some of the key events and individuals involved.

This paper describes in situ meteorological forcing and evaluation data, and bias-corrected reanalysis forcing data, for cold regions' modelling at 10 sites. The long-term datasets (one maritime, one arctic, three boreal, and five mid-latitude alpine) are the reference sites chosen for evaluating models participating in the Earth System Model-Snow Model Intercomparison Project. Periods covered by the in situ data vary between 7 and 20 years of hourly meteorological data, with evaluation data (snow depth, snow water equivalent, albedo, soil temperature, and surface temperature) available at varying temporal intervals. Thirty-year (1980–2010) time series have been extracted from a global gridded surface meteorology dataset (Global Soil Wetness Project Phase 3) for the grid cells containing the reference sites, interpolated to 1 h time steps and bias-corrected. Although the correction was applied to all sites, it was most important for mountain sites hundreds of metres higher than the grid elevations and for which uncorrected air temperatures were too high and snowfall amounts too low. The discussion considers the importance of data sharing to the identification of errors and how the publication of these datasets contributes to good practice, consistency, and reproducibility in geosciences. The Supplement provides information on instrumentation, an estimate of the percentages of missing values, and gap-filling methods at each site. It is hoped that these datasets will be used as benchmarks for future model development and that their ease of use and availability will help model developers quantify model uncertainties and reduce model errors. The data are published in the repository PANGAEA and are available at https://doi.pangaea.de/10.1594/PANGAEA.897575.

The GPS refractometry method is presented to derive snow water equivalent (SWE) by using refracted GPS L1 signals from a low-cost antenna buried underneath the snowpack. The GPS monitoring system is installed at 2536ma.s.l. in the Swiss Alps. GPS refractometry is able to correct the influence of the snowpack above the buried antenna. The systematic and stochastic snow induced effects in the GPS residuals are significantly reduced by estimating the SWE above the antenna. The refined method is able to estimate the SWE hourly over a full season and results are highly correlated to the reference sensors data with a median relative bias of less than 10 %.

Originally developed to fulfill the needs of applications consuming meteorological data, the Open Source MeteoIO library offers many valuable features for operational systems. It has been from the very start developed with a focus on robustness and efficiency for applications that must run unattended. Its use has been expanded to cover a wide variety of operational tasks from monitoring to data standardization to data hub and from time series at single locations to time series of gridded data. It is now used by several operational services as well as several companies.

This paper describes ESM-SnowMIP, an international coordinated modelling effort to evaluate current snow schemes, including snow schemes that are included in Earth system models, in a wide variety of settings against local and global observations. The project aims to identify crucial processes and characteristics that need to be improved in snow models in the context of local- and global-scale modelling. A further objective of ESM-SnowMIP is to better quantify snow-related feedbacks in the Earth system. Although it is not part of the sixth phase of the Coupled Model Intercomparison Project (CMIP6), ESM-SnowMIP is tightly linked to the CMIP6-endorsed Land Surface, Snow and Soil Moisture Model Intercomparison (LS3MIP).

The application of physically-based numerical modeling of the snowpack in support of avalanche hazard prediction is increasing. Modeling, in complement to direct observations and weather forecasting, provides information otherwise unavailable on the present and future state of the snowpack and its mechanical stability. However, there is a significant mismatch between the capabilities of modeling tools developed by research organizations and implemented by some operational services, and the actual operational use of those by avalanche forecasters, thereby causing frustration on both sides. By summarizing currently implemented modeling tools specifically designed for avalanche forecasting, we intend to diminish and contribute to bridge this gap. We highlight specific features and potential added value, as well as challenges preventing a more widespread use of these modeling tools. Lessons learned from currently used methods are explored and provided, as well as prospects for the future, including a list of the most critical issues to be addressed.

The energy balance of an alpine snow cover significantly changes once the snow cover gets patchy. The local advection of warm air causes above-average snow ablation rates at the upwind edge of the snow patch. As lateral transport processes are typically not considered in models describing surface exchange, e.g., for hydrological or meteorological applications, small-scale variations in snow ablation rates are not resolved. The overall model error in the hydrological model Alpine3D is split into a contribution from the pure "leading edge effect" and a contribution from an increase in the mean air temperature due to a positive snow-albedo feedback mechanism. We found an overall model error for the entire ablation period of 4% for the almost flat alpine test site Gletschboden and 14% for the Wannengrat area, which is located in highly complex terrain including slopes of different aspects. Terrestrial laser scanning measurements at the Gletschboden test site were used to estimate the pure "leading edge effect" and reveal an increase in snow ablation rates of 25–30% at the upwind edge of a snow patch and a total of 4–6% on a catchment scale for two different ablation days with a snow cover fraction lower than 50%. The estimated increase of local snow ablation rates is then around 1–3% for an entire ablation period for the Gletschboden test site and approximately 4% for the Wannengrat test site. Our results show that the contribution of lateral heat advection is smaller than typical uncertainties in snow melt modeling due to uncertainties in boundary layer parameters but increases in regions with smaller snow patch sizes and long-lasting patchy snow covers in the ablation period. We introduce a new temperature footprint approach, which reproduces a 15% increase of snow ablation rates at the upwind edge of the snow patch, whereas observations indicate that this value is as large as 25%. This conceptual model approach could be used in hydrological models. In addition to improved snow ablation rates, the footprint model better represents snow mask maps and turbulent sensible heat fluxes from eddy-covariance measurements.

Surface snow in polar and mountainous regions is often mobile and this mobility influences surface mass balance and isotopic composition before final deposition, which is poorly understood thus far. In December 2016 and January 2017, during a field campaign in Queen Maud Land, Antarctica, a snowfall and subsequent drifting snow events were recorded by meteorological and drifting snow stations. Associated small-scale topography changes and snow hardness changes were measured by terrestrial laser scanning and with a SnowMicroPen. The polar field measurements show that drifting snow is necessary for wind packing and thereby confirm previous findings from wind tunnel experiments. However, the snow hardness in Antarctica was significantly higher than what was achieved in the wind tunnel. This is most likely due to higher wind speeds and more intense saltation in the natural environment. As in the wind tunnel, no influence of time at a scale of days was observed on the hardness. This suggests that time and sintering are not the dominating processes in wind packing but that the impact compaction at the moment of deposition may be more important. Furthermore, it is quantitatively demonstrated how the reorganization of fresh snow into barchan dunes during subsequent drifting snow events is accompanied by significant increases in surface hardness at all locations on the dune. However, with the available data, the hardness variability on the dune could not be explained satisfactorily. In particular and unlike in the wind tunnel, there was no correlation between the hardness and the wind exposure parameter Sx. This is most likely because the measurements of hardness and the wind exposure situation were not simultaneous. This shows that highly temporally resolved snow depth data are necessary to investigate wind packing in more detail. These results form an important step in understanding how drifting snow links precipitation to deposition via snow hardening.

Wind sometimes creates a hard, wind-packed layer at the surface of a snowpack. The formation of such wind crusts was observed during wind tunnel experiments with combined SnowMicroPen and Microsoft Kinect sensors. The former provides the hardness of new and wind-packed snow and the latter spatial snow depth data in the test section. Previous experiments had shown that saltation is necessary but not sufficient for wind-packing. The combination of hardness and snow depth data now allows to study the case with saltation in more detail. The Kinect data requires complex processing but with the appropriate corrections, snow depth changes can be measured with an accuracy of about 1 mm. The Kinect is therefore well suited to quantify erosion and deposition. We found that no hardening occurred during erosion and that a wind crust may or may not form when snow is deposited. Deposition is more efficient at hardening snow in wind-exposed than in wind-sheltered areas. The snow hardness increased more on the windward side of artificial obstacles placed in the wind tunnel. Similarly, the snow was harder in positions with a low Sx parameter. Sx describes how wind-sheltered (high Sx) or wind-exposed (low Sx) a position is and was calculated based on the Kinect data. The correlation between Sx and snow hardness was −0.63. We also found a negative correlation of −0.4 between the snow hardness and the deposition rate. Slowly deposited snow is harder than a rapidly growing accumulation. Sx and the deposition rate together explain about half of the observed variability of snow hardness.

Global Navigation Satellite Systems (GNSS) contribute to various Earth observation applications. The present study investigates the potential and limitations of the Global Positioning System (GPS) to estimate in situ water equivalents of the snow cover (snow water equivalent, SWE) by using buried GPS antennas. GPS-derived SWE is estimated over three seasons (2015/16–2017/18) at a high Alpine test site in Switzerland. Results are validated against state-of-the-art reference sensors: snow scale, snow pillow, and manual observations. SWE is estimated with a high correspondence to the reference sensors for all three seasons. Results agree with a median relative bias below 10% and are highly correlated to the mean of the three reference sensors. The sensitivity of the SWE quantification is assessed for different GPS ambiguity resolution techniques, as the results strongly depend on the GPS processing.

Snow is a key element for many socioeconomic activities in mountainous regions. Due to the sensitivity of the snow cover to variations of temperature and precipitation, major changes caused by climate change are expected to happen. We analyze the evolution of some key snow indices under future climatic conditions. Ten downscaled and postprocessed climate scenarios from the ENSEMBLES database have been used to feed the physics-based snow model SNOWPACK. The projected snow cover has been calculated for 11 stations representing the diverse climates found in Switzerland. For the first time, such a setup is used to reveal changes in frequently applied snow indices and their implications on various socioeconomic sectors. Toward the end of the twenty-first century, a continuous snow cover is likely only guaranteed at high elevations above 2000 m a.s.l., whereas at mid elevations (1000–1700 m a.s.l.), roughly 50 % of all winters might be characterized by an ephemeral snow cover. Low elevations (below 500 m a.s.l.) are projected to experience only 2 days with snowfall per year and show the strongest relative reductions in mean winter snow depth of around 90 %. The range of the mean relative reductions of the snow indices is dominated by uncertainties from different GCM-RCM projections and amounts to approximately 30 %. Despite these uncertainties, all snow indices show a clear decrease in all scenario periods and the relative reductions increase toward lower elevations. These strong reductions can serve as a basis for policy makers in the fields of tourism, ecology, and hydropower.

Wind-packed snow in the form of slabs or crusts is an important part of alpine and polar snow covers. Yet, the formation process of such layers is poorly understood. For example, it remains unclear whether drifting snow is necessary for wind-packing or not. A better understanding of wind-packing could improve snow-cover models and avalanche danger forecasts and contribute to the assessment of mass balances in polar regions. We designed a closed-circuit, obround wind tunnel to study the process of wind crust formation. A SnowMicroPen was used to measure how the hardness of the snow evolved. The results show that no crust forms without saltation. Drifting snow is a necessary but not sufficient condition for wind-packing. The dynamics of erosion and deposition appear to be equally important.