Snow cover is a key component in polar regions and, besides the ocean, is the second largest interface between the atmosphere and the Earth’s surface during winter (Valt and Salvatori, 2016). The annual snow layer is critical to the surface mass balance of the glaciers, as it is the result of snow accumulation and ablation, with the former controlled by the snowfall rate and the latter depending on the net incoming energy flux at the surface (Kumar et al., 2019). On a larger scale, owing to its high reflectivity in the visible light range and to the significant mass of water stored in it, seasonal snow cover also has a strong impact on regional and global energy and water cycles. In addition, due to its high specific surface area, snow is capable of absorbing and scavenging a wide variety of natural and anthropogenic gases, particles, and contaminants from the atmosphere, which are eventually preserved within the seasonal snowpack until summer snowmelt (Gallet et al., 2011; Spolaor et al., 2014; Vecchiato et al., 2018; Burgay et al., 2021). Snowpacks also constitute habitats for precipitated microorganisms that actively interact with their physico-chemical environment and are driven by ecosystem changes (Els et al., 2019).

In the Arctic Archipelago of Svalbard, the seasonal snowpack covers ∼60 to 100 percent of the land between winter and summer (Gallet et al., 2019). From a physical point of view, snow deposited on land or ice surfaces is thermodynamically unstable and very sensitive to changes both in weather and/or climate conditions. In fact, the snowpack is an extraordinarily dynamic medium that can vary throughout the season from non-isothermal to isothermal conditions (Singh, 2011). During winter, it is usually composed of a succession of snow layers with different densities depending on the processes by which they have been subjected (i.e., wind-compacted layers, melt-refrozen layer, ice layer, bottom hoar). This condition leads to a non-homogeneous thermal profile of the snowpack, with temperatures generally lower at the snow-air interface and higher at the base of the snowpack (Fierz, 2011). Such a thermal gradient promotes the metamorphosis of snow crystals, resulting in recrystallization and faceting within the snowpack (Pinzer and Schneebeli, 2009; Pinzer et al., 2012). Another process that can significantly alter the chemo-physical properties of the snowpack frequently occurs in autumn and winter when, due to the presence of a semi-continuous weather front between the cold masses of Arctic air and the warmer air of the polar cell (Urbański and Litwicka, 2022), Svalbard weather is often subjected to rapid changes. Incursions of southerly mild moist air can lead to significant atmospheric warming (with a relative increase up to 20°C–25°C), shifting temperatures from well below zero values to positive, in a matter of a few days. Such short-lived intrusions of warm spells can carry aerosols originating in the mid-latitudes and moisture (Spolaor et al., 2021), causing rainfall events commonly identified as “rain on snow events” (ROS) (Serreze et al., 2021). In a warmer climate scenario, the ROS are predicted to increase during winter linked with the weakening of the polar vortex, warmer sea waters, and the disappearance of sea ice (Bintanja and Selten, 2014; Bintanja and Andry, 2017; Müller et al., 2022). Finally, in response to increasing temperatures and insolation in late spring and summer, especially in the coastal region, the snowpack is subject to melting, becoming isothermal and homogeneous in stratification.

Over the last decades, due to positive-feedback processes involving several components of the Earth system (Goosse et al., 2018), the Arctic region is experiencing an enhancement of near-surface air temperature 3 to 4 times more compared to lower latitudes and the global mean: a process widely known as Arctic Amplification (Maturilli et al., 2013; Previdi et al., 2021; Rantanen et al., 2022). In this perspective, a deeper understanding of the physical snowpack evolution in response to the ongoing atmospheric warming is critical to better evaluate and model future interactions between a changing snow cover and other Earth’s compartments, including atmosphere, glaciers, soils, freshwater networks, ocean, and sea ice (Gallet et al., 2019). Moreover, continuous observations of snow depth and physical properties carried out in parallel to atmospheric aerosol observations at the nearby stations of Gruvebadet and Zeppelin provide new opportunities to progress in the understanding of the potential effect of black carbon and dust deposition on snow albedo and metamorphism (AMAP, 2015). Finally, the understanding of melt water penetration in the snowpack and internal refreezing also needs to be improved due to the complexity of the processes involved and the lack of field observations (D’Amboise et al., 2017).

Manual measurements of the main snow physical properties (e.g., temperature, density, snow stratigraphy, layer harness, type of crystals, specific surface area) are usually carried out directly in the field through the snow pit technique, which consists of digging a trench in the annual snowpack and acquiring the snow physical parameters at different depths (Gallet et al., 2019). This technique is the most widely used and valuable to compare measurements collected in different sites, in order to assess the spatial variability of the snowpack conditions. The snow pit approach, however, requires the snowpack to be altered or disrupted, it is time-consuming and provides only a snapshot of a broader annual variability of the snow cover. As a valid alternative to field work measurements, a new generation of non-invasive field devices installed on near-surface-based automatic nivometric stations, has given the possibility to continuously measure different snowpack physical properties, such as snow depth, snow temperature profile, snow surface temperature, water liquid content, snow water equivalent and many others (Kinar and Pomeroy, 2015). To date, automatic observations of the snow cover are regularly conducted by snow and avalanche services primarily in mountain areas to provide the public with timely information about the snow condition and avalanche hazards (Pirazzini et al., 2018). The data obtained from snow monitoring, although punctual, are crucial and lead to an improvement in physical and chemical models of the snowpack, providing high-resolution inputs to numerical climate models, hydrology, snowpack evolution (Kinar and Pomeroy, 2015; Pirazzini et al., 2018). In the polar regions, instead, long-term automatic measurements of the snowpack for scientific purpose are relatively scarce and limited to few study cases due to severe weather conditions and extremely challenging logistics (Pfeffer and Humphrey, 1996; Järvinen et al., 2013; Gallet et al., 2019; Domine et al., 2021; Royer et al., 2021).

In late 2020, to fill this observational gap and assist in the interpretation of the distribution of chemical species in the snowpack, an automated nivometric station (ANS) was installed in Ny-Ålesund in West Spitzbergen, Svalbard. It provides continuous measurements of snow depth (SD), snow thermal profile, liquid water content (LWC) and fractional snow-cover area (fSCA). In this paper, we present the ANS instrumental set-up, the first-year record of preliminary measurements collected together with manual observations for comparison, the main issues encountered and the possible future implementations to fully assess the temporal evolution of the snowpack in real time.

Located along the coast of the Kongsfjord, on the western side of Svalbard, Ny-Ålesund is the world’s second northernmost year-round research station hosting research projects and long-term observation series. In November 2020, the first instrumental set-up of the automated nivometric station was installed on a flat area over the tundra in the proximity of the Gruvebadet Aerosol Laboratory, ∼1 km Southwest from Ny-Ålesund (78° 55.018’ N, 11° 53.698’ E, 43 m a.s.l., Figure 1). A 2-m high triangular truss, constituting the supporting structure of the ANS, was fixed to the ground through steel cables and anchor rods, adjacent to a delimited area dedicated for the year-round monitoring of snow chemistry since 2018.

The original instrumental set-up includes an ultrasonic snow-meter sensor measuring snow depth (model NESA-LU06, Resolution = 1mm, Accuracy <0.1% full-scale value at 20°C), 6 Resistance Temperature Detectors (RTD) Pt100 to measure the snowpack thermal profile (model NESA-TT, Sensitivity = 0.01°C, Accuracy ≤0.1°C) and 4 frequency domain reflectometry (FDR) soil moisture sensors to detect the amount of liquid water content within the snowpack (model NESA-SM1, Sensitivity = 0.1%, Precision = ±2%). In addition, in April 2021, the existing ANS infrastructure was implement with a near-infrared (NIR) time lapse-camera to calculate the fractional snow-cover area and observe specific processes occurring at the snow surface (i.e., roughness, wind crust formations, snow metamorphism, etc.). The NIR camera is equipped with a Sony IMX219 sensor (8 Megapixel resolution), installed in a heated housing mounted on top of the ANS truss, and coupled to an IR light source. The sensor was modified by removing an IR cut filter and adding a Storaro blue NIR filter. This last filter allows the removal of the red colour component and the detection of the near-infrared light component. The camera was programmed to shoot an oblique picture of the snow surface every hour (Figure 1). To obtain an accurate measure of the snow temperature at different snow depths and minimise the disturbance of the ANS structure on the free settlement of the snowpack, we designed and built a separate support unit on which temperature sensors were installed. The separate unit is composed of 6 Pt100 thermistors supported by vertical PVC pipes with different heights (5, 25, 50, 75, 100, 125 cm above the ground level), fixed to a wooden base ∼2 m far from the ANS supporting truss (Figure 1). Snow temperature measurements are obtained only by sensors that were completely buried by snow. To assess whether a given sensor was effectively surrounded by snow, we combine automatic snow depth measurements with time-lapse imagery from the NIR camera. Temperatures provided by sensors not completely buried by snow were discarded. Moreover, to record the LWC variations at different snow depths, we installed 4 Frequency Domain Reflectometry (FDR) moisture sensors on the extremity of their supporting arms at 25, 50, 75, 100 cm above the ground, respectively. The supporting arms were distributed radially around the ANS truss with an outreach of about 50 cm, to reduce the physical interference among them (especially during the melting season) and ensure that each sensor is fully immersed in the snow at its corresponding level. All the sensors were calibrated by their manufacturing companies, checked before installation, and connected to a datalogger (model NESA-Evolution) for continuous acquisition. Snow physical properties are measured every 10-s and logged with 10-min time resolution. The datalogger was powered by 220V A/C electrical supply (with a 12V back-up battery) and connected to the internet via Ethernet. Data transfer is assured by FTP connection through the local network. More detailed information about data coverage, sampling time resolution and other technical specifications of the snow monitoring system installed on the ANS, is provided in Table 1.

In the following sections, we present the first-year measurements of snow depth, snow temperature and LWC profiles acquired in Ny-Ålesund during the snow season between November 2020 and July 2021. We complement our observational record with measures of the fSCA derived from NIR time lapse imagery starting in late spring and lasting until the end of the snow melting season (April-June 2021). We also show manual measurements of snow depth and snow temperature profiles carried out at weekly resolution for comparison.

Although the snow physical data acquired by the ANS during the first year of snow monitoring in Ny-Ålesund showed a high potential for future developments, we would highlight some technical issues encountered during the snow season to suggest some possible solutions. First, we point out that, especially after heavy rainfall, the snowpack level sensor can convey water underneath, creating a melting cone and altering the snow height measurement. Although this problem cannot be physically avoided, a possible solution is to compare the SD measured by the ultrasonic sensor with the SD data inferred from the NIR camera images to assess possible bias between the two positions.

Secondly, we did not have information about snow temperature between the snow surface and the highest temperature sensor not discarded (as shown in Figure 2). These data gaps can be filled by employing an infrared snow surface temperature sensor. This improvement will allow for a more comprehensive picture of the snowpack thermal gradient, from the bottom to the surface. In this perspective, measuring near-surface atmospheric temperature directly at the ANS site could provide more accurate details about the air-snow surface interactions from both a physical and a chemical point of view. Third, although a direct measure of the bulk density of the snowpack can be obtained only by manual measurements, it can be indirectly inferred from the combined automatic measurements of snow water equivalent (SWE) (e.g., using a snow pillow, a neutron probe, or a Snowpack Analyser) and snow depth (Pirazzini et al., 2018). Since the ANS is already equipped with an ultrasonic snow level sensor, we foresee coupling it with a SWE sensor to automatically retrieve snow density variations during the snow season. This will enhance the harmonisation of the measurements with other snow observational study sites, in line with the indications provided by the Global Cryosphere Watch Initiative for improving in-situ snow depth and SWE observations (Brun et al., 2013). Last but not least, at the beginning of polar night, we reported damages to the separate unit of the ANS holding the snow temperature probes, due to the presence of a reindeer. Future developments of the ANS will include a fence to shield the instruments from local fauna and vice versa.

This work aims to present the instrumental set-up composing our ground-based automated nivometric station (ANS), installed in November 2020 in Ny-Ålesund, Svalbard. In addition to the instrumental specification, we provide two records of automatic and manual measurements of snow depth and snow temperature together with a record of liquid water content in the snowpack acquired over the first year of snow monitoring (winter 2020 - summer 2021). Furthermore, in April 2021, we installed a near-infrared time lapse camera to measure the fractional snow-cover and observe surface snow features. The snow season lasted 225 days in total, between November ‘20 and June ‘21, with a maximum snow depth of 117 cm observed in May 2021. During both winter and summer periods, we observed a strong connection between near-surface atmospheric temperature and the snow temperatures in the upper 50 cm of the snowpack. In contrast, lower snow layers were less affected by external temperature changes at least on diurnal time scales. The combined information from snow temperature and LWC data indicate four distinct melting events. Three melting events, associated with rain-on-snow events, occurred in a matter of a few days, and only impacted the basal part of the snowpack. The last melting event, corresponding to the summer snow melt, involved the whole snowpack and lasted ∼40 days, until the seasonal snow cover completely disappeared in the first days of July 2021. In this regard, measurements of the fractional snow-cover from the NIR camera indicated that the snow cover lasted 4 days more compared with automatic snow depth measurements by the ultrasonic sensor. In view of the foreseen future developments, including the installation of new sensors for measuring snow surface temperature, near-surface air temperature, and snow water equivalent (SWE), the ANS is the first automated, comprehensive snowpack monitoring system in Ny-Ålesund measuring key snow physical properties. The data collected are fundamental to a deeper knowledge of the snowpack evolution over the snow season and its response to warming climate and/or increased predicted extreme weather events such as rain on snow events. In addition, continuous records of physical snow observations could be extremely valuable to constrain and further develop numerical simulation of the seasonal snowpack with a focus on the distribution and total load of contaminants and impurities as well as to better understand the microbiological activity within it.