ID36: Monitoring of heterogeneous mountain snowpacks

Abstract ID 792 | Date: 2022-09-13 10:10 – 10:20 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

D’Amboise, Christopher; Fromm, Reinhard; Adams, Marc; Ploerer, Matthias; Demmler, Christian; Teich, Michaela

Topography and total roughness of the terrain, which is composed of surface roughness (approximately 1 m-scale) and terrain roughness (approximately 5-10 m-scale), influence the distribution of snow as well as snowpack properties. Down woody debris such as logs, stumps, branches or root plates as well as rocks and other obstacles in the terrain cause surface roughness. These features disturb the snowpack and can influence avalanche formation by creating an inhomogeneous and noncontinuous snow stratigraphy and by anchoring the snowpack. The uneven distribution/redistribution of snow due to wind will promote inhomogeneous snow depths and snow stratigraphy, which can disrupt the continuity of a potential weak layer or sliding surface. Roughness elements distributed along the fall line can therefore help anchoring the snowpack to the ground as a snow supporting structure. By accounting for the roughness in the direction of the fall line of the terrain, it might be possible to further refine the delineation of potential avalanche release areas.

We present a newly developed automated directional roughness algorithm using a point cloud surface model (PCSM). The algorithm defines a roughness score for each point contained in the PCSM by approaching obstacles from different rotational directions (0°-180°). The algorithm attempts to separate the terrain, similar to a point cloud terrain model (PCTM, which is compiled from the rotational direction investigation), from the roughness of the PCSM. The roughness of the terrain is quantified by using the differences between the PCSM and the PCTM. The slope of the smoothed terrain is approximated as a byproduct of the rotational directional roughness calculation and can be used to investigate the capability for roughness to act as snow supporting structures.

We applied our algorithm to an area with a wind-disturbed forest with many stumps, root plates and logs laying on and above the ground providing surface roughness and calculated the roughness along the automatically detected fall line. Additionally, the direction of the maximum and minimum roughness can be extracted. The PCSM of the study area was derived from UAV-based structure-from-motion photogrammetry. With our tool, we can now investigate how the height, distribution and orientation of down woody debris affect the protective effect of a disturbed forest with regards to potential avalanche release areas. In general, directional roughness can be computed in any direction with this algorithm, be it the terrain fall line or wind direction, which has implications for quantifying the effect of surface roughness on snow distribution/redistribution.

Abstract ID 504 | Date: 2022-09-13 10:20 – 10:30 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Valence, Eole (1,2); Mckenzie, Jeffrey (2); Charonnat, Bastien (1); Baraer, Michel (1)

Due to climate change, the mountain cryosphere is rapidly evolving. The most perceptible changes outside of the polar regions occur in mountainous areas, which host a significant amount of the cryosphere. In alpine area, the retreat of the cryosphere is even more pronounced as atmospheric warming increases with the altitude. Thus, change in the cryosphere affect the hydrology of these areas.

Our research focuses on the hydrological behaviour of the Grizzly Creek, a subarctic catchment in the Kluane Mountains, and in particular understanding the behaviour of seasonal snow distribution. The catchment is in the south-west of the Yukon at an elevation of around 1500 m.a.s.l. and is composed by diverse cryospheric elements, including bare glaciers, debris-covered glaciers, rock glaciers, permafrost, and seasonal snow cover. However, this system has no significant surface water discharge, indicating potential groundwater flow out of the catchment. Thus, identify and quantify the hydrological pathways and fluxes in this type of watershed is crucial to face climate changes in arctic and subarctic mountainous regions. To quantify the amount of groundwater flow, identify it sources is necessary. Thus, the distribution of snow cover is of great interest in such catchment.

We present a multi-method approach to maps the seasonal snow cover repartition, including satellite imagery, ground-based lidar measurements, snow depth and snow water equivalent monitoring, UAV-based photogrammetry, and time-lapse imagery. In-situ monitoring is used to classify snow amount for several points used as references. The satellite imagery permits to transpose the in-situ observation over the entire catchment.

The results of the research provide an improved understanding of snow distribution in northern mountain catchment, with implications for improving hydrologic models and forecasting.

Abstract ID 298 | Date: 2022-09-13 10:30 – 10:40 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Bühler, Yves (1,2); Stoffel, Andreas (1,2); Marty, Mauro (3); Eberhard, Lucie (1,2); Bührle, Leon (1,2)

Snow depth, its distribution and the amount of water stored in the snowpack are key parameters for a wide range of applications such as snow avalanche mitigation, hydropower generation, and ecological questions. Despite its crucial importance, operational snow depth measurements are sparse and limited to point locations (e.g., automated weather stations and observations). However, thus cannot capture the high spatial variability of the snowpack, and therefore considerably limiting its significance.

Recent studies proofed the ability of photogrammetric methods to map spatially continuous snow depth over large regions with very high spatial resolution. Around Davos, Switzerland we acquired, terrestrial, drone based, airborne and spaceborne (very high-resolution satellites – VHR) photogrammetric measurements since several years. We achieve snow depth accuracies with a root mean square error (RMSE) in the range of a few cm (UAS) and 0.15 m (piloted airplane) to around 0.5 m (satellites). In this talk we give insights in the performed snow depth mapping campaigns and an overview on the results and their validation. Then we demonstrate how these datasets can be applied in practice, for example to monitor avalanche mitigation measures, to evaluate potential locations for automated weather stations or to document extreme avalanche events. During the past years, photogrammetric snow depth mapping developed from a very experimental state to a key tool for snow and avalanche research at SLF.

Abstract ID 122 | Date: 2022-09-13 10:40 – 10:50 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Vickers, Hannah (1); Malnes, Eirik (1); Eckerstorfer, Markus (2)

The characteristics of snow cover are highly sensitive to variations in temperature and precipitation. In Svalbard these are undergoing significant change in response to a rapidly warming climate and the associated positive feedback processes. The occurrence of wintertime rain-on-snow (ROS) events are expected to increase in frequency and intensity across the Arctic as a result of climate change. ROS events dramatically alter snow cover characteristics, by saturating the snowpack and enhancing surface runoff as well as causing widespread formation of ground ice, which can negatively impact many ecosystems as well as infrastructure. Knowledge of the spatial and temporal variations in ROS occurrence across Svalbard, both past and present is needed to understand which areas are most vulnerable to ROS hazards and how this may change in the future. This work has utilised Synthetic Aperture Radar (SAR) observations to produce an 18-year dataset of wet snow cover observations for Svalbard, from which a method for detecting and mapping both spring melt onset and ROS frequency has been developed. The mean spatial variations in melt onset and ROS occurrence reflect the geographical gradients in temperature and precipitation across the archipelago and are largely in agreement with those reproduced by downscaling of output from regional climate models. The timing of ROS onset as detected using the SAR observations coincide well with in-situ measurements of rainfall, however in some cases the duration of a ROS event cannot be reliably estimated using SAR observations of wet snow, in particular after phase transitions from rain to snow. Linear trends derived from the limited time series of observations suggests that ROS frequency is increasing over most of the archipelago, with largest increases in the south and along the western coast. However, low elevation areas in the central parts of the archipelago exhibit a decreasing trend in ROS over the time period of observations.

Abstract ID 426 | Date: 2022-09-13 10:50 – 11:00 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Schauwecker, Simone; Ayala, Álvaro; Cortés, Gonzalo; Goubanova, Katerina; Macdonell, Shelley

Melting snow from high-elevation areas dominates discharge and freshwater supply in semi-arid Chile. The snow cover largely depends on few winter events and consequently there is a large year-to-year variability in the snow water equivalent (SWE). The extraordinarily dry conditions experienced almost continuously since 2010 and increased water consumption in the region have led to a considerable stress of the water system. For an efficient water allocation and water management, it is therefore crucial to know the actual SWE stored in the mountain snowpack. Until now, decisions are based on point measurements of the SWE or snow area estimations from MODIS. A drawback of these estimations are the large uncertainties that hamper an efficient water allocation with important implications for water security of different areas such as hydropower, agriculture and domestic use.

In this project we are developing a new operational SWE Estimation Tool for water resources decision-making in the Coquimbo region (SWEET-Coquimbo). The SWE is estimated using a data assimilation framework that combines meteorological forcing ensembles from reanalysis data, hydrological modelling and satellite observations of the snow-covered area. First, meteorological data are disaggregated to the model grid resolution and a comparison with station data from a relatively dense network of meteorological stations allows evaluating the bias and uncertainty of reanalysis data. This uncertainty is then used to generate ensembles of input variables for the hydrological model. The data assimilation further assigns higher weights to ensemble replicates that generate best predictions compared to satellite observations. These weights are used to calculate ensemble metrics, such as the mean and uncertainty of SWE estimates. Finally, the model outputs are validated with in-situ manual SWE measurements and UAV-based and LiDAR observations of the snow depth.

Along with the operational SWE estimates, retrospective SWE estimates over 1985-present will be used to improve our understanding of the spatial distribution and temporal evolution of the SWE in the region. The results will be published on an open access web platform and available as inputs for downstream applications such as hydrological models and hydrological assessments.

Abstract ID 572 | Date: 2022-09-13 11:00 – 11:10 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Deschamps-Berger, César (1,2,3); Gascoin, Simon (2); Dumont, Marie (3); Berthier, Etienne (4); Luks, Bartłomiej (5); López-Moreno, Juan Ignacio (1); Lafaysse, Matthieu (3); Shaw, Thomas (6); Brun, Fanny (7); Koch, Franziska (8); Gallet, Jean-Charles (9); Revuelto, Jesus (1); Haddjeri, Ange (3)

Developing methods to map snow depth at high resolution (<10 m) is an active field of research as snow depth is a key variable for water resource and avalanche risk assessment. Close range remote sensing is commonly used, combining lidar or photogrammetry with an airplane or a drone. However, drones acquisition are limited to small basins (<10 km²) and airborne campaigns require logistics hard to meet in many mountains of the world. Recent improvements in satellite photogrammetry provide an alternative to map the snow depth in any place of the world at high spatial resolution (~3 m) by differencing digital elevation models with and without snow derived from satellite stereoscopic images.

Here, we present approaches to monitor the high spatial variability of the snowpack by using a collection of snow depth maps calculated from images of the Pléiades satellite in the Alpes, the Andes, the Himalayas, the Pyrenees, the Sierra Nevada (USA) and Svalbard. First, the comparison with a reference snow depth map measured with airborne lidar in the Sierra Nevada (140 km²) provides a robust estimation of the Pléiades snow depth error. At the 3 m pixel scale, the standard error is about 0.7 m. The error decreases to 0.3 m when the snow-depth maps are averaged over areas greater than 10³ m². With this accuracy, Pléiades snow-depth maps allow the observation of the impact of the processes shaping mountain snowpack (wind transport, avalanches) and the description of the high spatial variability of the snow depth. Then, we describe statistically the structure of the snow depth distribution with semi-variograms and relate it to the various climates, relief and processes of the study sites. Finally, a multi-annual time series of snow depth maps in the Pyrenees is assimilated in a distributed snowpack model. The assimilation corrects precipitation bias in the meteorological forcing and the lack of spatial variability of the modeled snowpack.

Abstract ID 483 | Date: 2022-09-13 11:10 – 11:20 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Nagler, Thomas (1); Schwaizer, Gabriele (1); Barella, Riccardo (2); Bueso-Bello, Jose-Luis (3); Cremonese, Eduardo (4); Essery, Richard (5); Hetzenecker, Markus (1); Jonas, Tobias (6); Keuris, Lars (1); Koch, Roland (7); Marin, Carlo (2); Mölg, Nico (1); Notarnicola, Claudia (2); Olefs, Marc (7); Rizzoli, Paola (3); Rott, Helmut (1); Santos, Benedita (2); Strasser, Ulrich (8); Warscher, Michael (8); Fitrzyk, Magdalena (9); Volden, Espen (9)

The seasonal snow cover is an important resource in mountain regions, the monitoring of which is crucial for water management, snow hydrology, tourism, and natural hazard mitigation. Alpine snow is also an important source of headwater for drainage basins downstream. AlpSnow (2020-2022) is a science activity within ESA’s Alpine Regional Initiative, addressing the development of novel Earth Observation techniques and algorithms for the generation of innovative snow products optimized for specific scientific and operational applications. The AlpSnow portfolio includes products on snow extent, snow mass, wetness, albedo, and snow surface grain size.

Algorithm development and validation activities are supported by field activities in five test areas in different Alpine regions. Each test area is equipped with automated stations providing time series of meteorological and snow measurements. These are supplemented by dedicated field activities in order to collect additional snow reference data for validation. In the first phase of the project several candidate algorithms for each snow parameter were selected, implemented, tested, and evaluated. Based on the intercomparisons we selected a set of preferred algorithms which are further optimized for complex topography and diverse surface cover types. Regarding high resolution snow extent, we use an advanced linear multispectral unmixing approach and a new method applying machine learning techniques, both exploiting the spectral capabilities of Sentinel-2 and Landsat. Different algorithms are also tested and evaluated for surface albedo and grain size products. Wet snow extent, snow water equivalent (SWE) and snow depth are derived from synthetic aperture radar (SAR) data. For SWE two approaches are explored. The first is based on the assimilation of EO snow extent into the physical SNOWGRID model of the Austrian meteorological service (ZAMG), the second approach studies the use of repeat-pass SAR interferometry for generating maps of snow accumulation in mountain regions at about hundred meter grid size. This algorithm is tested using L-Band SAR data from the ALOS-2 PALSAR and SAOCOM missions and prepares for the future Copernicus Expansion Mission ROSE-L. For mapping the snow depth during the snowmelt season we investigate the suitability of differencing DEMs from the TanDEM-X mission acquired on snow-free surfaces and over wet snow. The impact of the products will be assessed by six use cases dealing with snow modelling, hydrology, and water management.

We will present results of the experiments dedicated to the selection of retrieval algorithms and report on the properties and performance of the prototype algorithms and products.

Abstract ID 672 | Date: 2022-09-13 11:20 – 11:30 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Bair, Edward (1); Stillinger, Timbo (1); Rittger, Karl (2); Abegg, Steph (2); Kleiber, Will (2)

Because of the heterogeneity of mountain snowpacks, there is a need for daily snow cover mapping at the slope scale (≤ 30 m) that is unmet for a variety of scientific users, ranging from: hydrologists, to the military, to wildlife biologists. One reason this goal is challenging is because it cannot be fulfilled by any single sensor. Recently, constellations of satellites and fusion techniques have made noteworthy progress. Here, two such recent advances are examined: 1) a fused MODIS – Landsat 8 OLI product with daily 30 m spatial resolution; and 2) a harmonized Landsat 8 – Sentinel 2A/B product with 2-3 day temporal and 30 m spatial resolution. State-of-art spectral unmixing techniques are applied to surface reflectance products from 1 & 2 to create snow cover and albedo maps. Then an energy balance model is run to reconstruct snow water equivalent (SWE). As a baseline for comparison, previous SWE reconstructions using snow cover from MODIS at 463 m daily resolution are used. For validation, lidar-based Airborne Snow Observatory SWE estimates are used. This work addresses questions about whether increasing resolution yields worthwhile increases in accuracy, or just adds more pixels and computing time. Such questions and answers are fundamental to the study of snow in mountains worldwide, and will inform future scientific priorities and mission specifications.

Abstract ID 312 | Date: 2022-09-13 11:30 – 11:40 | Type: Oral Presentation | Place: SOWI – Seminar room U3 |

Rittger, Karl (1); Skiles, Mckenzie (2); Musselman, Keith (1); Serreze, Mark (3); Brodzik, Mary J. (3)

Water managers need accurate observations of snow cover and albedo to make decisions for a diverse set of applications. Remotely sensed snow cover and albedo products that are currently available do not meet operational requirements for several reasons. First, the most widely available snow products from MODIS use an index algorithm developed in 1989 that employs two spectral bands, one in a visible wavelength and one in the shortwave-infrared, which together provide limited information to estimate fractional coverage or snow albedo. Second, cloud cover is difficult to discriminate from snow and causes data gaps that can be filled using techniques that rely on accurate cloud masking. Third, MODIS is an elderly system that continues to operate well beyond its design life, and the successor to MODIS, the Visible Infrared Imaging Radiometer Suite (VIIRS), do not gap-fill the snow cover product and there is no snow albedo product. Fourth, standard snow cover and snow albedo products do not account for off-nadir observations that introduce uncertainty and additional data gaps. Fifth, no widely distributed product accounts for the darkening of snow caused by light absorbing particles and their impact on snow albedo.

We have partnered with snow remote sensing end users who serve diverse needs of national and international water resource decision makers. With their guidance, we create and provide daily gap-filled snow cover and snow albedo, including impacts of light absorbing particles. The products account for off-nadir views, snow under the forest canopy, and use cloud filtering techniques not employed in existing products. Using algorithms shown to perform consistently across sensors—specifically Landsat 8/9, MODIS, and VIIRS—we will process the historical daily record and produce data in near real-time with a sub-daily latency period for both MODIS Terra and VIIRS Suomi. This project will complete the transition of the data processing, archiving, and distribution to the National Snow and Ice Data Center. The transition will ensure the continued production of snow cover and snow albedo products for the lifetime of these sensors.

Our partner organizations are decision makers poised to directly benefit from accurate and timely snow cover and snow albedo information. These collaborators are globally distributed in North America, New Zealand, the Andes, High-Mountain Asia, and the European Alps. The historical and near real-time snow and albedo products will enhance decision-making processes to better inform stakeholders in a range of applications, including streamflow forecasting, agriculture, and water futures planning.

Abstract ID 772 | Date: 2022-09-13 15:15 – 15:17 | Type: Poster Presentation | Place: SOWI – Garden |

Radlherr, Alexander; Winkler, Michael

The snowpack is a key component in several fields like climatology, hydrology, or natural hazards research and mitigation, not least in mountainous regions. One of the most considerable snowpack features is the snow water equivalent (SWE), representing the mass of water stored in the snowpack and – in another perspective – the weight straining objects the snow is settling on (snow load). In comparison to other snow properties, like e.g. snow depth, SWE is rather complex to measure and consecutive observations do not have a long tradition in many regions.

Despite various recent developments in measuring SWE by means of remote sensing or other noninvasive methods, e.g. with pressure sensors, scales, GNSS sensors, cosmic-ray neutron probes etc., the standard measuring technique still is using snow tubes or gauging cylinders, often in combination with digging pits. The cylinder or tube designs very a lot: from meters long metal coring tubes of typical inner diameters of ca. 4-7 cm (without the need of pits) or PVC cylinders with typical lengths of 0.5 to 1.5 m and diameters ranging from about 5-20 cm to small aluminum tubes holding a maximum of 0.5 liter of snow. Comparison and calibration experiments traditionally use (one of) these “standard methods” as reference. However, studies addressing their accuracy, precision and repeatability are rare.

This contribution provides first results of several field tests at different sites in the Austrian Alps covering a great variety of snow conditions (e.g. dry and wet), snow depths and SWEs, respectively. Different types of SWE measurement tubes are compared to each other but they are also confronted with “absolute” observations. For the latter 3 x 4 m rectangular areas have been cleared of snow and the respective snow masses have been weighed stepwise using ~50 liter buckets. Known issues like increasing accuracy with increasing diameters are confirmed, however, many statistical measures like variance and bias vary quite a lot depending on the equipment used. Furthermore, a synopsis of the suitability of the various methods depending on the problem or the objective of the observation is provided.

Abstract ID 521 | Date: 2022-09-13 15:17 – 15:19 | Type: Poster Presentation | Place: SOWI – Garden |

Kollert, Andreas; Mayr, Andreas; Rutzinger, Martin

Obtaining retrospective information on snow covered area (SC) by means of remote sensing is challenged by a mismatch of desired temporal and spatial resolution and the characteristics of available image collections. Oftentimes, coarse resolution (CR) SC products (e.g. derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) or Advanced Very-High-Resolution Radiometer) do not meet the requirements of mountain researchers who focus on processes taking place at a significantly smaller scale. Especially in high-mountain regions, such CR products fail to adequately represent the heterogeneity of the complex terrain. Still, there are large archives of CR (e.g. MODIS) products that provide a temporal coverage that is superior to other archives (e.g., Landsat). Since there is some temporal overlap of CR satellite image archives with medium to high resolution (MHR) satellite image archives (Landsat, Sentinel-2), MHR products are increasingly exploited to downscale CR products. We present first results of downscaling MODIS snow products to Sentinel-2 resolution (i.e. 20 m) by an implementation of methodologically different, existing approaches (probabilistic, Machine Learning) and compare their performance at a high alpine study site in the Stubai Alps (Tyrol, Austria). Emphasis is put on challenges that arise when working in high-mountain areas, namely the performance and errors of the different methods with respect to the complex terrain (i.e. topography, surface roughness and vegetation cover). Methods developed will be validated for robustness and transferability ensuring applicability in high-alpine study sites with similar characteristics.

Abstract ID 973 | Date: 2022-09-13 15:21 – 15:23 | Type: Poster Presentation | Place: SOWI – Garden |

Fierz, C. (1); Marty, Christoph (1); Salzmann, N. (1,2); Nitu, R. (3); De Rosnay, P. (4); Luojus, K. (5); Roulet, Y.-A. (6)

Proposal for a WMO Snow Monitoring Competence Centre
C. Fierz 1
, N. Salzmann 1,2
, C. Marty1 , R. Nitu 3
, P. de Rosnay 4
, K. Luojus 5
, Y.-A. Roulet 6
1WSL institute for Snow and Avalanche Research SLF, Switzerland
2Climate Change, Extremes and Natural Hazards in Alpine Regions Research Center CERC
3World Meteorological Organization, Switzerland
4European Centre for Medium-Range Weather Forecasts, United Kingdom
5Finnish Meteorological Institute, Finland
6Federal Office of Meteorology and Climatology MeteoSwiss, Switzerland
Snow as frozen precipitation is globally of increasing importance in a world
that, on the one hand, faces more frequent droughts, where snow and ice
can play an important role as water storage; on the other hand, it is
confronted with more extreme precipitation events, where snow can
dampen immediate run-off but also cause avalanches or floods. Moreover,
decreasing snow cover due to climate warming and due to increasing dust
and soot loads lowers the planetary albedo, which changes the energy
balance of our planet. For all these reasons, the global monitoring of
different snow variables is of increasing significance.
In recent years, the idea of a Centre on Competence in Snow Monitoring
(CCiSM) came up within the Global Cryosphere Watch (GCW) community.
First ideas were presented at scientific conferences and WMO workshops
and meetings. A concept with a concrete ‘business plan’ has now been
finalised, resulting in a proposal submitted to the WMO Infrastructure
Commission (INFCOM). The proposal is to build on the existing mature
framework of WMO Measurement Lead Centres and to propose establishing
knowledge-based competence hubs that would assume functions in support
of sustaining the quality of snow observations and the quality of data and
including capacity development activities. Additionally, these hubs would
link to the GCW data portal to facilitate the access to data sets and data
providers. Once the framework is established, one or more centres could
apply to obtain this designation
The proposal of these centres it timely, since this coincides with the launch
of the newly established Joint Body on the Status of the Mountain Snow
Cover (JB-SMSC), a joint venture between the International Association of
Cryospheric Sciences (IACS), the Mountain Research Initiative (MRI), and
GCW. A strong mutual benefit is expected, including the proposed centres
as long-term repositories for the knowledge developed within the scope of
JB-SMSC.
It is proposed that Snow Monitoring Competence Centres become
specialised WMO Measurement Lead Centre providing, for example,
answers to questions concerning in situ measurement practices of common
snow variables, standards for snow data quality checks, as well as
connecting both measurement and modelling experts to raise awareness of
each other needs.
The poster will outline the benefits of establishing such centres