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.