The European Space Agency (ESA) Global Snow Monitoring for Climate Research (GlobSnow) snow water equivalent (SWE) v2.0 data record contains snow information derived for the Northern Hemisphere since 1979 to present day. The snow water equivalent describes the amount of liquid water in the snow pack that would be formed if the snow pack was completely melted. The data record is produced using a combination of passive microwave radiometer and ground-based weather station data. The SWE record is produced on a daily, weekly, and monthly basis. SWE information is provided for terrestrial non-mountainous regions of the Northern Hemisphere, excluding glaciers and ice sheets. The data are provided in HDF4 and NetCDF formats along with PNG browser images for quick viewing. A single file contains the data for a single day and contains two fields: the SWE estimate and an error estimate (standard deviation).

The GlobSnow SWE record, based on methodology by Pulliainen (Pulliainen 2006 and Takala et al. 2011), utilizes a data-assimilation based approach combining space-borne passive radiometer data (SMMR, SSM/I, and SSMIS) with data from ground-based synoptic weather stations. The satellite sensors utilized provide data at K- and Ka-bands (19 GHz and 37 GHz, respectively) at a spatial resolution of approximately 25 km. The SWE product is projected to Equal-Area Scalable Earth Grid (EASE-Grid) and provides the daily SWE estimates for whole Northern Hemisphere (lambert's equal-area azimuthal projection) in a single file. Although the EASE-Grid projection can represent data almost to the equator, the product is limited between latitudes 35° and 85° for physical reasons (extent of seasonal snow cover). The input satellite data for the SWE prototype products are from SMMR, SSM/I, and SSMIS sensors acquired from National Snow and Ice Data Center, Boulder Colorado, U.S.A. (NSIDC).

The snow water equivalent product is based on the combination of satellite-based microwave radiometer and ground-based weather station data.

The long term SWE data set spans from 1979 – to present day The SWE maps are produced as daily, weekly, and monthly composites Nimbus-7 SMMR, DMSP (F8/F11/F13) SSM/I, and DMSP F17 SSMIS are the main data sources: SMMR for 1979 - 1987

ESA GlobSnow v3.0 snow water equivalent (SWE) dataset. Time series of 1980-2018 Northern Hemisphere terrestrial (non-mountainous) snow water equivalent data, containing monthly and monthly-bias-corrected SWE estimates.

Dataset constructed by combining satellite-based passive microwave radiometer data (Nimbus-7 SMMR, DMSP 5D2 SSM/I and DMSP 5D3 SSMIS) with ground based synoptic snow depth observations using bayesian data assimilation, incorporating HUT Snow Emission model; by Pulliainen et al. 1999 (doi:10.1109/36.763302) and Lemmetyinen et al. 2011 (doi:10.1109/TGRS.2010.2041357). […]

Datasets of daily snow water equivalent (SWE) over the Northern Hemisphere (excluding Greenland) were constructed using a multi-dataset approach for the time period 1981-2020. The data are on a regular 0.5-degree grid, with a threshold maximum of 2000mm. The general methodology for the creation of these datasets follows that of Mudryk et al. (2015). Supplemental Information Monthly snow cover fraction (SCF) and monthly snow water equivalent (SWE) are calculated using daily SWE data taken from the following four sources over the 35-year period from 1981 to 2016. Four sources of daily SWE data: The Modern-Era Retrospective Analysis for Research and Applications Version 2 (MERRA-2) (Global Modeling and Assimilation Office, 2016 [Ref 1]; Gelaro et al., 2017 [Ref 2] is a National Aeronautics and Space Administration (NASA) atmospheric reanalysis product generated with the Goddard Earth Observing System Model, Version 5.2.0 (GEOS-5), atmospheric general circulation model and atmospheric data assimilation system (ADAS). The temperature index model described by Brown et al. (2003) [Ref 3] reconstructs daily SWE using 6-hourly temperature field and 12-hourly precipitation field inputs from ERA-Interim reanalysis. This simplified index model includes most of the temperature-dependent processes included in the snow component of numerical land surface schemes (e.g. partitioning of precipitation into solid and liquid fractions, melt from rain-on-snow events, specification of new snowfall density, snow aging, and snowmelt). The physical snowpack model Crocus simulates daily SWE using meteorology from ERA-Interim (Brun et al., 2013) [Ref 4]. The European Space Agency GlobSnow product (Version 2; www.globsnow.info, last access: 20 September 2016) is a gridded product derived through a combination of satellite passive microwave data, forward snow emission model simulations, and climate station observations for non-alpine regions of the Northern Hemisphere (Takala et al., 2011) [Ref 5]. The GlobSnow product is partially masked over mountainous regions, defined as regions with a slope of 2 degrees or larger. SWE was replaced in grid cells which contain mountains with a blend of the GlobSnow SWE data (if any) and the mean value from the other three data sources. The weighting for the blend was determined by the fraction of the grid cell area which is mountainous. For grid cells with no mountainous terrain, unaltered GlobSnow data are used. As the fraction of mountainous terrain increases, the weight applied to the GlobSnow data is linearly reduced, reaching zero for grid cells containing only mountainous terrain. For a given dataset of daily SWE, the data was interpolated to a regular 0.25 degree grid over Canada. Monthly SCF is produced by applying a 4mm threshold to each of the four daily SWE fields to produce a daily binary snow cover field; this daily field is averaged over each month to produce a monthly snow cover fraction and the four data sets are averaged together. Monthly SWE is produced by averaging the regridded daily SWE fields from each source over the given month. Annual maximum SWE fields are calculated as the maximum value of daily SWE attained at each grid location over a given snow season. Note: grid cell values for SCF, SWE and SWEmax have been weighted by the fraction of land surface within each grid cell (excluding ocean, lakes and glaciers/ice caps). For example, a grid cell containing 50% land and 50% ocean which is fully snow-covered for the month will have a listed SCF value of 50%. Ref 1. Global Modeling and Assimilation Office: MERRA-2tavgM_2d_slv_Nx: 2d, Monthly mean, Time-Averaged, Single-Level, Assimilation, Single-Level Diagnostics V5.12.4, Greenbelt, MD, USA, Goddard Earth Sciences Data and Information Services Center (GESDISC), https://doi.org/10.5067/AP1B0BA5PD2K, last access: 11 April 2017b. Ref 2. Gelaro, R., McCarty, W., Suárez, M. J., Todling, R., Molod, A.,Takacs, L., Randles, C. A., Darmenov, A., Bosilovich, M. G., Reichle, R., Wargan, K., Coy, L., Cullather, R., Draper, C., Akella,S., Buchard, V., Conaty, A., da Silva, A. M., Gu, W., Kim, G.,Koster, R., Lucchesi, R., Merkova, D., Nielsen, J. E., Partyka,G., Pawson, S., Putman, W., Rienecker, M., Schubert, S. D.,Sienkiewicz, M., and Zhao, B.: The Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2),J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1, 2017. Ref 3. Brown, R., Brasnett, B., and Robinson, D.: Gridded North American monthly snow depth and snow water equivalent for GCM evaluation, Atmos.-Ocean, 41, 1–14, 2003. Ref 4. Brun, E., Vionnet, V., Boone, A., Decharme, B., Peings, Y., Valette,R., Karbou, F., and Morin, S.: Simulation of Northern Eurasian local snow depth, mass, and density using a detailed snowpack model and meteorological reanalyses, J. Hydrometeorol., 14,203–219, https://doi.org/10.1175/JHM-D-12-012.1, 2013. Ref 5. Takala, M., Luojus, K., Pulliainen, J., Derksen, C., Lemme-tyinen, J., Kärnä, J.-P., and Koskinen, J.: Estimating northern hemisphere snow water equivalent for climate research through assimilation of space-borne radiometer data and ground-based measurements, Proc. Spie., 115, 3517–3529,https://doi.org/10.1016/j.rse.2011.08.014, 2011.

The ESA funded GlobSnow project produced snow water equivalent (SWE) daily standard errors (Variance estimates) for the Northern Hemisphere for the years 1979-2013. SWE describes the amount of liquid water in the snow pack that would be formed if the snow pack was completely melted. The SWE product shall cover the Northern Hemisphere, excluding the mountainous areas, Greenland, the glaciers and snow on ice (lakes/seas/oceans). The spatial resolution of the product is 25 km on EASE-grid projection. Construction of the 30 years historical data set will be carried out using SMMR, SSM/I and SSMI/S data along with ground-based weather station data. The data are utilized for the different years as follows: 1979/09/11 - 1987/10/30 SMMR (Scanning Multichannel Microwave Radiometer onboard Nimbus-7 satellite) 1987/11/01 - 2008/12/31 SSM/I (Special Sensor Microwave/Imager onboard the DMSP satellite series F8/F11/F13) 2009/01/01 - present SSM/I(S) (Special Sensor Microwave/Imager (Sounder) onboard the DMSP satellite series F17/F18/) These data may be redistributed and used without restriction.

The European Space Agency (ESA) Global Snow Monitoring for Climate Research (GlobSnow) snow water equivalent (SWE) v2.0 data record contains snow information derived for the Northern Hemisphere since 1979 to present day. The snow water equivalent describes the amount of liquid water in the snow pack that would be formed if the snow pack was completely melted. The data record is produced using a combination of passive microwave radiometer and ground-based weather station data. The SWE record is produced on a daily, weekly, and monthly basis. SWE information is provided for terrestrial non-mountainous regions of the Northern Hemisphere, excluding glaciers and ice sheets. The data are provided in HDF4 and NetCDF formats along with PNG browser images for quick viewing. A single file contains the data for a single day and contains two fields: the SWE estimate and an error estimate (standard deviation).

The GlobSnow SWE record, based on methodology by Pulliainen (Pulliainen 2006 and Takala et al. 2011), utilizes a data-assimilation based approach combining space-borne passive radiometer data (SMMR, SSM/I, and SSMIS) with data from ground-based synoptic weather stations. The satellite sensors utilized provide data at K- and Ka-bands (19 GHz and 37 GHz, respectively) at a spatial resolution of approximately 25 km. The SWE product is projected to Equal-Area Scalable Earth Grid (EASE-Grid) and provides the daily SWE estimates for whole Northern Hemisphere (lambert's equal-area azimuthal projection) in a single file. Although the EASE-Grid projection can represent data almost to the equator, the product is limited between latitudes 35° and 85° for physical reasons (extent of seasonal snow cover). The input satellite data for the SWE prototype products are from SMMR, SSM/I, and SSMIS sensors acquired from National Snow and Ice Data Center, Boulder Colorado, U.S.A. (NSIDC).

The snow water equivalent product is based on the combination of satellite-based microwave radiometer and ground-based weather station data.

• The long term SWE data set spans from 1979 – to present day
• The SWE maps are produced as daily, weekly, and monthly composites
• Nimbus-7 SMMR, DMSP (F8/F11/F13) SSM/I, and DMSP F17 SSMIS are the main data sources: SMMR for 1979 - 1987; SSM/I for 1987 - 2009, SSMIS for 2010 - present
• Coverage: Non-mountainous of Northern Hemisphere, excluding glaciers and ice sheets
• The SWE maps are generated using EASE-Grid projection and a 25km spatial resolution

More detailed descriptions are found within the project documentation accessible through the project Web pages.

The ESA funded GlobSnow project produced snow water equivalent (SWE) 7-day estimates for the Northern Hemisphere for the years 1979-2013.

SWE describes the amount of liquid water in the snow pack that would be formed if the snow pack was completely melted. Weekly Aggregated Snow Water Equivalent (Weekly L3B SWE) were calculated for each day based on a 7-day sliding time window aggregation of the daily SWE product.

The SWE product shall cover the Northern Hemisphere, excluding the mountainous areas, Greenland, the glaciers and snow on ice (lakes/seas/oceans)

The spatial resolution of the product is 25 km on EASE-grid projection.

Construction of the 30 years historical data set will be carried out using SMMR, SSM/I and SSMI/S data along with ground-based weather station data. The data are utilized for the different years as follows:

1979/09/11 - 1987/10/30 SMMR (Scanning Multichannel Microwave Radiometer onboard Nimbus-7 satellite) 1987/11/01 - 2008/12/31 SSM/I (Special Sensor Microwave/Imager onboard the DMSP satellite series F8/F11/F13) 2009/01/01 - present SSM/I(S) (Special Sensor Microwave/Imager onboard the DMSP satellite series F17/F18/)

These data may be redistributed and used without restriction.

This data set is a daily gridded terrestrial snow water equivalent (SWE) dataset based on five component SWE products:

• GlobSnow combined SWE product (passive mircowave/ground-based weather station, version 2)
• ERA-Interim/Land reanalysis SWE product
• MERRA reanalysis SWE product
• Crocus SWE data set: output from the Crocus snowpack model, driven by ERA-Interim meteorology (Brun et al. 2013)
• GLDAS SWE product (version 2) (Rodell et al. 2004; Rodell and Beaudoing 2013)

This dataset contains key characteristics about the data described in the Data Descriptor GlobSnow v3.0 Northern Hemisphere snow water equivalent dataset. Contents:

    1. human readable metadata summary table in CSV format


    2. machine readable metadata file in JSON format

The efforts of the European Space Agency (ESA) Data User Element (DUE) funded GlobSnow project has resulted in two new hemispherical records of snow parameters intended for climate research purposes. The dataset contains satellite-retrieved information on snow water equivalent (SWE) extending 34 years. The record on snow water equivalent is produced using a combination of passive microwave radiometer and ground-based weather station data, spanning years 1979 to 2013. The GlobSnow SWE record, based on methodology by Pulliainen (Pulliainen 2006, Takala et al. 2011) utilizes a data-assimilation based approach combining space-borne passive radiometer data (SMMR, SSM/I and SSMIS) with data from ground-based synoptic weather stations. The satellite sensors utilized provide data at K- and Ka-bands (19 GigaHertz and 37 GigaHertz respectively) at a spatial resolution of approximately 25 kilometers (km). The SWE record is produced on a daily, weekly and monthly basis. SWE information is provided for terrestrial non-mountainous regions of Northern Hemisphere, excluding glaciers and Greenland.

This paper presents an analysis of observed and simulated historical snow cover extent and snow mass, along with future snow cover projections from models participating in the 6th phase of the World Climate Research Programme Coupled Model Inter-comparison Project (CMIP6). Where appropriate, the CMIP6 output is compared to CMIP5 results in order to assess progress (or absence thereof) between successive model generations. An ensemble of six observation-based products is used to produce a new time series of historical Northern Hemisphere snow extent anomalies and trends; a subset of four of these products is used for snow mass. Trends in snow extent over 1981-2018 are negative in all months, and exceed -50 x 103 km2 during November, December, March, and May. Snow mass trends are approximately -5 Gt/year or more for all months from December to May. Overall, the CMIP6 multi-model ensemble better represents the snow extent climatology over the 1981-2014 period for all months, correcting a low bias in CMIP5. Simulated snow extent and snow mass trends over the 1981-2014 period are stronger in CMIP6 than in CMIP5, although large inter-model spread remains in the simulated trends for both variables. There is a single linear relationship between projected spring snow extent and global surface air temperature (GSAT) changes, which is valid across all scenarios. This finding suggests that Northern Hemisphere spring snow extent will decrease by about 8% relative to the 1995-2014 level per °C of GSAT increase. The sensitivity of snow to temperature forcing largely explains the absence of any climate change pathway dependency, similar to other fast response components of the cryosphere such as sea ice and near surface permafrost. Supplemental Information The data in these four files are described in Mudryk et al. (2020). monthly NH snow mass over the 1981-2018 time period based on estimates from MERRA2, Crocus, Brown, and GlobSnow. monthly NH snow extent over the 1967-2018 time period based on estimates from NOAA Snow Chart record, JASMES, MERRA2, Crocus, Brown, and GlobSnow. March snow extent based on Brown (2000) recalibrated to match 2. April snow extent based on Brown (2000) recalibrated to match 2. A full description of the analysis is contained in Mudryk et al. 2020 References: Brown, R. D.: Northern Hemisphere snow cover variability and change, 1915–1997, J. Climate, 13, 2339–2355, 2000. Mudryk, L., Santolaria-Otín, M., Krinner, G., Ménégoz, M., Derksen, C., Brutel-Vuilmet, C., Brady, M., and Essery, R.: Historical Northern Hemisphere snow cover trends and projected changes in the CMIP-6 multi-model ensemble, The Cryosphere Discuss., https://doi.org/10.5194/tc-2019-320, in review, 2020.

High Asia is a high-altitude region in Asia dominated by the Qinghai-Tibet Plateau. It is an important distribution area for snow in the middle and low latitudes. The dynamic changes of snow cover have an important impact on the water and energy cycle, climate and environmental changes in the region. Snow-covered snow water equivalent (SWE) products (such as ESA GlobSnow), which are widely used around the world, have data vacancies in high Asia, and some global algorithms are generally overestimated in high Asia. Based on the AMSR-E brightness temperature data, this data is based on the NASA standard algorithm and the improved algorithm of the Qinghai-Tibet Plateau with different underlying surfaces. Two sets of ASMR-E snow-water equivalent data sets from 2002 to 2011 in the Qinghai-Tibet Plateau were produced. The improved algorithm is verified and evaluated by the measured snow depth of the ground meteorological station, indicating that the accuracy of the improved algorithm of the Qinghai-Tibet Plateau has improved. In addition, the paper also comprehensively processes the SWE products based on MODIS daily cloud-free snow cover products and NASA snow algorithm and obtains a set of comprehensive data sets of microwave and optical snow water equivalents from 2003 to 2011. The three sets of snow water equivalent product data sets provide basic data support for scientific research in climate change, water, energy balance, environmental change and water use in high Asia.

Supported by the Strategic Priority Research Program of the Chinese Academy of Science (XDA19070100). Tao Che, the director of this program, who comes from Key Laboratory of Remote Sensing of Gansu Province, Northwest Institute of Eco-Environment and Resources, CAS. They used machine learning methods combined with multi-source gridded snow depth product data to derive a long-time series over the Northern Hemisphere. Firstly, the applicability of artificial neural network (ANN), support vector machine (SVM) and random forest (RF) method in snow depth fusion are compared. It is found that random forest method shows strong advantages in snow depth data fusion. Secondly, using the random forest method, combined with remote sensing snow depth products such as AMSR-E, AMSR-2, NHSD and GlobSnow and reanalysis data such as ERA-Interim and MERRA-2. These gridded snow depth products and environmental factor variables are used as the input independent variables of the model. In situ observations of China Meteorological Station (945), Russia Meteorological Station (620), Russian snow survey data (514), and global historical meteorological network (41261) are used as reference truth to train and verify the model. The daily gridded snow depth dataset of the snow hydrological year from 1980 to 2019 (September 1 of the previous year to May 31 of the current year) is prepared on the cloud platform provided by the CASEarth. Since the passive microwave brightness temperature data from 1980 to 1987 is the data of every other day, there will be a small number of missing trips in the data during this period. Using the ESM-SnowMIP and independent ground observation data for verification, the quality of the fusion data set has been improved. According to the comparison between the ground observation data and the snow depth products before fusion, the determination coefficient (R2) of the fusion data is increased from 0.23 (GlobSnow snow depth product) to 0.81, and the corresponding root mean square error (RMSE) and mean absolute error (MAE) are also reduced to 7.7 cm and 2.7 cm.

This dataset contains Daily Snow Cover Fraction of viewable snow from AATSR, produced by the Snow project of the ESA Climate Change Initiative programme.

Snow cover fraction viewable (SCFV) indicates the area of snow viewable from space over all land surfaces. In forested areas this refers to snow viewable on top of the forest canopy. The SCFV is given in percentage (%) per pixel.

The global SCFV product is available at 0.01° grid size (about 1 km) for all land areas, excluding Antarctica and Greenland ice sheet. The coastal zones of Greenland are included. The SCFV time series provides daily products for the period 2002 – 2012.

The SCFV product is based on Advanced Along-Track Scanning Radiometer (AATSR) data aboard the Envisat satellite.

The retrieval method of the snow_cci SCFV product from AATSR data has been further developed and improved based on the ESA GlobSnow approach (Metsämäki et al. 2015) and complemented with a pre-classification module. In a first step, clouds are masked using an adapted version of the Simple Cloud Detection Algorithm version 2.0 (SCDA2.0) (Metsämäki et al. 2015), defined as SCDA2.3. All cloud-free pixels are then used for the snow extent mapping, using spectral bands centred at about 659 nm and 1.61 µm, and an emissive band centred at about 10.85 µm. The snow_cci snow cover mapping algorithm is a two-step approach: first, a strict pre-classification is applied to identify all cloud-free pixels which are clearly snow free. For all remaining pixels, the snow_cci SCFV retrieval method is applied. Improvements to the GlobSnow algorithm implemented for snow_cci version 1 include adaptation of the retrieval method for mapping in forested areas the SCFV.

Permanent snow and ice, and water areas are masked based on the Land Cover CCI data set of the year 2000. Both classes were separately aggregated to the grid size of the SCFV product. Water areas are masked if more than 30% of the grid cell is classified as water, permanent snow and ice areas are masked if more than 50% is identified as such areas in the aggregated map. The product uncertainty for observed land areas is provided as unbiased root mean square error (RMSE) per grid cell in the ancillary variable.

The SCFV product aims to serve the needs of users working with the cryosphere and climate research and monitoring activities, including the assessment of variability and trends, climate modelling and aspects of hydrology, meteorology, and biology.

The Norwegian Computing Center (Norsk Regnesentral, NR) is responsible for the SCFV product development and generation from AATSR data. The Remote Sensing Research Group of the University of Bern supported the development. ENVEO IT GmbH developed and prepared all auxiliary data sets used for the product generation.

There are a few days without any AATSR acquisitions in the years 2002, 2003, 2004, 2006, 2008, 2010 and 2012.

This dataset contains Daily Snow Cover Fraction (snow on ground) from MODIS, produced by the Snow project of the ESA Climate Change Initiative programme.

Snow cover fraction on ground (SCFG) indicates the area of snow observed from space on land surfaces, in forested areas corrected for the transmissivity of the forest canopy. The SCFG is given in percentage (%) per pixel.

The global SCFG product is available at about 1 km pixel size for all land areas, excluding Antarctica and Greenland ice sheets. The coastal zones of Greenland are included.

The SCFG time series provides daily products for the period 2000 – 2019.

The SCFG product is based on Moderate resolution Imaging Spectroradiometer (MODIS) data on-board the Terra satellite.

The retrieval method of the snow_cci SCFG product from MODIS data has been further developed and improved based on the ESA GlobSnow approach described by Metsämäki et al. (2015) and complemented with a pre-classification module developed by ENVEO. For the SCFG product generation from MODIS, multiple reflective and emissive spectral bands are used. In a first step, clouds are masked using an adapted version of the Simple Cloud Detection Algorithm version 2.0 (SCDA2.0) (Metsämäki et al., 2015). All cloud free pixels are then used for the snow extent mapping, using spectral bands centred at about 550 nm and 1.6 µm, and an emissive band centred at about 11 µm. The snow_cci snow cover mapping algorithm is a two-step approach: first, a strict pre-classification is applied to identify all cloud free pixels which are certainly snow free. For all remaining pixels, the snow_cci SCFG retrieval method is applied. Improvements to the GlobSnow algorithm implemented for snow_cci version 1 include (i) the utilisation of background and forest reflectance maps derived from statistical analyses of MODIS time series replacing the constant values for snow free ground and snow free forest used in the GlobSnow approach, and (ii) the usage of a global forest transmissivity map developed and created within snow_cci based on forest density from Hansen et al. (2013) and forest type layers from Land Cover CCI (Defourny, 2019). The forest transmissivity map is used to account for the shading effects of the forest canopy and estimate also in forested areas the fractional snow cover on ground.

Permanent snow and ice, and water areas are masked based on the Land Cover CCI data set of the year 2000. Both classes were separately aggregated to the pixel spacing of the SCFG product. Water areas are masked if more than 30 percent of the pixel is classified as water, permanent snow and ice areas are masked if more than 50 percent are identified as such areas in the aggregated map. The product uncertainty for observed land pixels is provided as unbiased root mean square error (RMSE) per pixel in the ancillary variable.

The SCFG product is aimed to serve the needs for users working in the cryosphere and climate research and monitoring activities, including the detection of variability and trends, climate modelling and aspects of hydrology, meteorology, and biology.

ENVEO is responsible for the SCFG product development and generation from MODIS data, SYKE supported the development.

There are a few days without any MODIS acquisitions in the years 2000, 2001, 2002, 2003, 2008, 2016 and 2018. On several days in the years 2000 to 2006, and on a few days in the years 2012, 2015 and 2016, the acquired MODIS data have either only limited coverage, or some of the MODIS data were corrupted during the download process. For these days, the SCFG products are available but have data gaps.

1. Data content: global 0.25-degree resolution daily snow depth product for seasonal snow from 2013 to 2020, including land areas in the northern and southern hemispheres except Antarctica and Greenland.
2. Data source and processing method: In view of the underestimation problem of current mainstream passive microwave snow products in mountainous areas, the machine learning algorithm is used to generate a new snow depth product, which trained the representative snow depth measurements from meteorological stations globally, year by year, based on both the AMSR2 (Advanced Microwave Scanning Radiometer 2) brightness temperatures and the prior knowledge of snow parameters from snow process model simulations. The snow priors come from the improved multiple-layer SNTHERM (Snow Thermal Model) simulation driven by the GLDAS (Global Land Data Assistance System) dataset, including snow grain size, snow density and the first-guess snow depth parameters. The machine learning algorithm is mainly used to overcome the biases in snow priors caused by the forcing dataset error. The meteorological station measurements participating in the construction of ground truths are from the GHCN (Global Historical Climate Network), GSOD (Global Surface Summary of the Day), USHCN (The United States Historical Climate Network), ECA&D (European Climate Assessment&Dataset) and daily snow depth measurements from reference meteorological stations in China. The snow cover extent is jointly determined based on MODIS 0.05 ° Climate Grid snow cover product and SNTHERM simulations.
3. Data quality description: 1) Cross validation: The root mean square error (RMSE) of retrieved snow depth validated by 1.72 million samples in eight years worldwide is 12.4 cm. The bias of snow depth is not sensitive to elevation and forest cover fraction. 2) Independent validation: the RMSE validated using the 25 km quadrat, 500 m quadrat and snow course measurements from the China Snow Survey in 2018-2019 years is 4.93-8.9 cm; RMSE validated against the EuroAlps dataset and CanSWE dataset are 17.3 cm and 24.2 cm, respectively, in the 0-100 cm snow depth range; if the snow depth is converted to the snow water equivalent (SWE) using a fixed density of 240 kg/m3, the SWE RMSE validated against the GlobSnow validation dataset - the Russian part and the US SNOTEL dataset are 29.2 mm and 40.1 mm in 0-200 mm range. Due to the differences between the training dataset and the independent validation dataset, it is difficult to have a fully-unbiased validation result in the deep snow range; however, the new product still slightly outperforms the JAXA official AMSR2 product and GlobSnow product in the mountains and at the edge of mountains, respectively. Note: The current validation results are calculated based on the representative in-situ measurements. Please contact the data provider and refer to the paper for representative evaluation criteria.
4. Data application achievements and prospects: research on the spatial and temporal distributions of seasonal snow cover; on requirement of a seasonal snow information to support the study of other elements in the cryosphere; researchers looking for a less-biased passive microwave snow depth product in the mountains at current stage.
5. Others: We would like to express our gratitude to the providers of all kinds of raw datasets to generate our new product. Please refer to the paper for the acknowledgement and data description.

This dataset contains Daily Snow Cover Fraction (snow on ground) from AATSR, produced by the Snow project of the ESA Climate Change Initiative programme.

Snow cover fraction on ground (SCFG) indicates the area of snow observed from space on land surfaces, in forested areas corrected for the transparency (“transmissivity”) of the forest canopy. The SCFG is given in percentage (%) per grid cell.

The global SCFG product is available at 0.01° grid size (about 1 km) for all land areas, excluding Antarctica and Greenland ice sheet. The coastal zones of Greenland are included.

The SCFG time series provides daily products for the period 2002 – 2012.

The SCFG product is based on Advanced Along-Track Scanning Radiometer (AATSR) data aboard the Envisat satellite.

The retrieval method of the snow_cci SCFG product from AATSR data has been further developed and improved based on the ESA GlobSnow approach (Metsämäki et al. 2015) and complemented with a pre-classification module. In a first step, clouds are masked using an adapted version of the Simple Cloud Detection Algorithm version 2.0 (SCDA2.0) (Metsämäki et al. 2015), defined as SCDA2.3. All cloud-free pixels are then used for the snow extent mapping, using spectral bands centred at about 659 nm and 1.61 µm, and an emissive band centred at about 10.85 µm. The snow_cci snow cover mapping algorithm is a two-step approach: first, a strict pre-classification is applied to identify all cloud-free pixels which are clearly snow free. For all remaining pixels, the snow_cci SCFG retrieval method is applied.

Improvements to the GlobSnow algorithm implemented for snow_cci version 1 include the usage of a global forest transmissivity map developed and created within snow_cci based on forest density from Hansen et al. (2013) and forest type layers from Land Cover CCI (Defourny 2019). The forest transmissivity map provides the local transparency of the forest canopy and is applied or estimating the fractional snow cover on the ground.

Permanent snow and ice, and water areas are masked based on the Land Cover CCI data set of the year 2000. Both classes were separately aggregated to the grid size of the SCFG product. Water areas are masked if more than 30% of the grid cell is classified as water, permanent snow and ice areas are masked if more than 50% is identified as such areas in the aggregated map. The product uncertainty for observed land areas is provided as unbiased root mean square error (RMSE) per grid cell in the ancillary variable.

The SCFG product aims to serve the needs of users working with the cryosphere and climate research and monitoring activities, including the assessment of variability and trends, climate modelling and aspects of hydrology, meteorology, and biology.

The Norwegian Computing Center (Norsk Regnesentral, NR) is responsible for the SCFG product development and generation from AATSR data. The Remote Sensing Research Group of the University of Bern supported the development. ENVEO IT GmbH developed and prepared all auxiliary data sets used for the product generation.

There are a few days without any AATSR acquisitions in the years 2002, 2003, 2004, 2006, 2008, 2010 and 2012.