Seasonal snow cover is an important component of the global climate system. It is highly variable in space and time and sensitive to both short term synoptic scale processes and long term climate-induced changes of hydrologic and atmospheric systems. The accurate quantification of the seasonal snow extent and mass is an important research issue to support climate analysis and modelling. Current snow products derived from different algorithms applied to various satellite data show significant discrepancies in extent and snow mass, a potential source for biases in climate monitoring and modelling. The recently launched ESA Climate Change Initiative (CCI+) Programme addresses seasonal snow as one of 9 Essential Climate Variables to be investigated for climate monitoring and modelling applications.

In the snow_cci project, scheduled for 2018 to 2021 in its first phase, reliable fully validated processing lines are developed and implemented. Based on these tools homogeneous multi-sensors time series for main parameters of the global seasonal snow cover are generated, with focus on the snow cover area extent and snow water equivalent (snow mass). Using GCOS guidelines, the requirements for these parameters are assessed and consolidated based on output of workshops and questionnaires addressing users dealing with different climate applications as well as CCI- projects related to other essential climate variables. The procedure for snow extent product generation includes algorithms for fractional snow extent retrieval and cloud screening, generating consistent daily products for snow on the surface (viewable snow) and snow on the surface corrected for forest masking (snow on ground) with global coverage. Input are medium resolution multi-spectral data from various sensors including AVHRR-2/3, AATSR, ATSR-2, MODIS, VIIRS, and SLSTR/OLCI covering the period from 1981 to present. An iterative development cycle is elaborated and implemented to allow for homogenisation of the snow extent products from different sensors by minimizing the bias. Independent validation of the snow products is performed using high resolution snow maps derived from Landsat and Sentinel-2 data acquired across different seasons, environments and climate zones around the globe from 1985 onwards as well as in-situ snow data following the community agreed Satellite Snow Product Validation and Inter-comparison Exercise (SnowPEx) validation.
Global time series of daily snow water equivalent (SWE) products are generated from passive microwave data from SMMR, SSM/I and AMSR from 1978 onwards. The GlobSnow SWE processor serves as the baseline version. It is further improved by updating the assimilation procedure which combines in-situ snow measurements and specific satellite data. Long-term stability and quality of the product is assessed using independent snow survey data and by intercomparison with the snow information from global land process models.
The usability of the snow_cci products is monitored by close interaction with the Climate Research Group, which performs case studies related to long term trends of seasonal snow extent and snow mass, evaluation of CMIP-6 and ESM SnowMIP simulations, and modelling of hydrological regimes in periods of climate change.
In this presentation we will summarize the requirements and product specification for the snow extent and SWE products with focus on climate applications and will present an overview of the algorithms and systems for generation of the time series. Demonstration products for daily global snow extent and snow water equivalent products from different sensors will be presented and first results of the multi-sensor consistency and validation activities as well as intercomparisons with snow products from other sources will be presented.

For much of the pre-melt and melt season, sunlight absorption can be the dominant melt energy source for snow and ice on land, glaciers, sea, and lakes. Snow and ice melt feedbacks play a role in Arctic amplification of climate warming.
The Copernicus ESA Sentinel-3 A and B platforms host two sensors with optical bands: the Ocean Land Color Instrument (OLCI) and the Sea and Land Surface Temperature Radiometer (SLSTR). These data enable production of snow/ice optical properties imagery for relatively flat areas. Retrieved snow and ice properties include:
snow extent
snow spectral albedo
snow broadband albedo
snow specific surface area
snow effective grain size
This presentation reviews data products, ground validation from the Antarctic and Greenland ice sheets and the French Alps and includes a discussion of a two-year (2017, 2018) climatological evaluation of Greenland ice sheet relative to past information from the NASA MODIS data overlapping with Sentinel-3A.

The seasonal snowpack, with its high albedo, low thermal conductivity and large water storing capacity, is a key component of the terrestrial energy, water, and carbon balance. At the same time, it is widely acknowledged that the accurate retrieval of the distribution of snow water equivalent (SWE) from Earth observing satellites remains an outstanding problem, especially in complex terrain. This is unfortunate because in terms of modulating the terrestrial balances, SWE is (along with albedo) arguably the most essential snow variable.

For decades, the depletion of snow-cover retrieved from optical sensors has been combined with snowmelt modeling to retrospectively reconstruct SWE in a deterministic manner. Despite some success, this approach ignores uncertainties in the retrievals, the meteorological forcing, and the snowmelt model, thus limiting its applicability. Recently, reanalyses based on ensemble smoothers that can account for these uncertainties have emerged as robust alternatives to traditional reconstruction. In this approach, the batch assimilation of the remotely sensed depletion of fractional snow-covered area (fSCA) is used to constrain an ensemble of modeled realizations of the seasonal snowpack. Such ensemble-based reanalyses (EBR) have already been used to estimate SWE at the hillslope scale (100 m) across entire mountain ranges with promising results.

The success of EBR hinges upon the quality of the remotely sensed fSCA depletion. We present results from an evaluation of fSCA retrieved from several optical satellite sensors that are currently in polar orbit, specifically: Sentinel-2A/2B MSI, Landsat 7 ETM+, Landsat 8 OLI and Terra/Aqua MODIS. For each of these sensors, we implemented and evaluated multiple retrieval algorithms ranging from simple thresholding of the normalized difference snow index to fully constrained spectral unmixing. For ‘ground truth’, we used high resolution (1-3 m) snow cover maps from two sites: the Bayelva catchment, near Ny-Ålesund (Svalbard, Norway), and the Mammoth Lakes basin (California, USA). At Bayelva, the reference maps were generated by classifying hundreds of orthorectified photographs taken between 2012 and 2017 from a time-lapse camera system overlooking the catchment from a nearby mountain peak. In the Lakes basin, maps were obtained from LIDAR measurements from multiple Airborne Snow Observatory flights in the 2016, 2017 and 2018 snow seasons. The results of this exercise reveal limitations in our ability to accurately infer fSCA without properly accounting for the mixed pixel problem, i.e. the fact that a pixel often consists of more than one endmember. This highlights the need for spectral unmixing even for the higher resolution sensors on-board the Sentinel-2 and Landsat constellations.

Examples are provided of how the Sentinel-2 constellation can help to constrain uncertainty in applications of EBR at the hillslope scale where previously only Landsat was (freely) available. We demonstrate how new global products such as the ERA5 meteorological reanalysis from the ECMWF offers hope for accurate EBR even in remote locations. Finally, we discuss the obstacles associated with generating an automated and global satellite-based EBR of the seasonal snowpack.

The European Alps are located close to one of the most industrialized areas of the planet and they are 3.000 km far from the largest desert of the Earth. Light-absorbing particles (LAPs) emitted from these sources can reach the Alpine chain and deposit on snow- and ice- covered areas and mountain glaciers. Furthermore, biological impurities (microorganisms and algae) can thrive on snow and glaciers and thus darkening their surfaces. Although several studies show that LAPs have important impacts on the optical properties of snow and ice, reducing the albedo and promoting the melt, the characterization of their impact in the European Alps is still in its infancy.
In this contribution, we present an overview of results of a multi-disciplinary approach aimed to study the impact of LAPs on snow and ice optical and thermal properties in the Alpine area. The impact of impurities on snow and ice has been analysed using satellite data (Landsat 8, Hyperion and Sentinel 2 data), field spectroscopy (with ASD and OceanOptics spectrometers), Automatic Weather Stations, aerial surveys (RGB and multispectral cameras on board unnmanned aerial vehicles), radiative transfer modeling (SNICAR, TARTES and BioSNICAR), snowpack modeling (CROCUS) and laboratory analysis (hyperspectral imaging system with HyIce). Furthermore, particle size (Coulter Counter), geochemical (Instrumental Neutron Activation Analysis, INAA), optical (Multi-Wavelength Absorbance Analyzer, MWAA), biological (genetic sequencing and algae density) analyses have been applied to determine the nature and radiative properties of particulate material deposited on snow and ice or aggregated into cryoconite holes.
Our results demonstrate that LAPs can be monitored from remote sensing at different scale in the Alps. LAPs showed to have a strong impact on the Alpine cryosphere, paving the way for assessing and predicting their role in melting processes.

The seasonal snow cover of the Northern Hemisphere (NH) is a crucial component of the global climate system. In the high latitudes, snow cover has a large influence on the hydrological cycle and many biogeochemical processes. Recent studies have shown negative trends in the extent and duration of the NH snow cover. Melt season timing and length estimates are important for both hydrological and climatological applications, such as spring flood forecasting and climate studies.

Melt season monitoring at continental scale is possible only from satellites, and the monitoring can be done by using both microwave and optical satellite methods. The optical satellite method uses changes in snow albedo to detect snow melt onset. When the melting starts, the albedo begins to decrease because of the snow metamorphism. The microwave satellite method, on the other hand, uses changes in microwave emissions of the snow cover. The microwave emissions depend on the physical properties of snow, which makes the method highly sensitive to the appearance of liquid water in the snowpack and therefore a very good tool for detecting snow melt onset.

In this study, we use two datasets of snow melt onset date (MOD): one from an optical satellite instrument and one from a microwave satellite instrument. The dataset from the optical satellite instrument is based on mean surface albedo from the Satellite Application Facility for Climate Monitoring (CM SAF) CLouds, Albedo and RAdiation second release Surface ALbedo (CLARA-A2 SAL) data (Anttila et al., 2018). The microwave-based dataset is produced by Environment and Climate Change Canada (ECCC) and the algorithm uses temporal variations in the differences of the brightness temperature (Wang et al., 2013). Our study covers land areas between latitudes 50°N and 80°N and period between 1982 and 2015. We use the nearest neighbor method to resample the datasets to the same projection and compare the melt onset dates pixel by pixel. To study trends in MOD, we use the Theil-Sen method and the Mann-Kendall test.

The analysis show that the microwave satellite method detects MOD about 10 days later than the optical satellite method. This results from the different methods the instruments use to detect snow melt onset: the albedo-based method detects the day, when snow metamorphism begins, whereas the microwave-based method detects the day, when meltwater appears in the snowpack. The difference in MOD between the datasets decreases in forests, where trees shade snow cover and therefore the snow metamorphism starts later compared to open areas. We studied trends in MOD across the NH as a whole and for Eurasia and North America separately, and in all cases the trend was negative and statistically significant.

Additionally, we will study mean temperatures on the melt onset date in different land cover classes. We will do this for both datasets separately to find out the mean temperatures on the day, when snow metamorphism begins and on the day, when meltwater appears in the snowpack. We will also compare all the results with in situ data and investigate, if they are consistent with each other.

References

Anttila, K., Manninen, T., Jääskeläinen, E., Riihelä, A., & Lahtinen, P. (2018). The Role of Climate and Land Use in the Changes in Surface Albedo Prior to Snow Melt and the Timing of Melt Season of Seasonal Snow in Northern Land Areas of 40° N–80° N during 1982–2015. Remote Sensing, 10(10), 1619.

Wang, L., C. Derksen, C., Brown, R., and Markus, T. (2013). Recent changes in pan-Arctic melt onset from satellite passive microwave measurements, Geophys. Res. Lett., 40, doi:10.1002/grl.50098.

The Greenland Ice Sheet (GrIS) is losing mass at a substantial and increasing rate, driven by increasing surface melt-induced runoff from its margins. The increasing mass loss is closely tied to a decrease in the reflectivity, or albedo, of the ice sheet's surface, particularly along its margins. Recent advances in the instrument intercalibration and surface albedo retrieval methods have allowed for the creation of a 34-year duration (1982-2015) timeseries of surface albedo data covering GrIS in its entirety during the summer months (MJJA). This dataset, called CLARA-A2-SAL (CM SAF Clouds, Albedo, and Radiation dataset, 2nd edition, Surface ALbedo), is used here to examine the trends and spatiotemporal variability in the reported albedo decline of GrIS during melt seasons.

The analysis is carried out based on the 5-day mean albedo data of CLARA-A2-SAL. Quality assurance procedures precede the analysis; any 5-day data based on an unreliably low number of clear-sky AVHRR overpasses is discarded. Also, an ice sheet mask is applied to ensure that the results are not contaminated by mixed signals from areas with fractional ice coverage (here, a threshold of 50% ice cover is applied). Also, any and all trend analyses are examined for robustness using a priori information on the dataset's retrieval stability over GrIS.

The actual analysis focuses on three main aspects. First, we divide the dataset temporally by month, examining each summer month's albedo changes separately, and further by decade, examining the differences in albedo changes between e.g. the 1980s and the 2000s.

Then, we examine the variability in the intensity of each melt season in the dataset period as well as changes in the albedo minima and maxima, indicative of e.g. intensifying melt or cumulative melt-induced emergence of previously deposited impurities along the ice sheet margins.

Finally, we examine the role of surface albedo changes in relation to the total ice mass lost in the study period, taking advantage of both state-of-the-art regional climate models and the latest gravimetric satellite observations of GrIS ice loss. The changes recorded by CLARA-A2-SAL are shown to be consistent with the reference sources as far back as the 1980s.

The results, which display an interesting range of spatiotemporal variability in the albedo change of the GrIS margins, are then compared against other recent satellite-based albedo datasets, displaying good conformity as far as dataset overlap allows, and discussed against the backdrop of literature regarding the ongoing changes of GrIS, and its drivers.