Based on the EUMETSAT AVHRR GAC FDR (Global Area Coverage), released in May 2023, a time series (1979–2022) was generated utilising the SCAmod algorithm of Metsämäki et al. (2015). Cloud masking relies on the probability mask of CLARA-A3, an upgrade of the existing cloud albedo and radiation (CLARA) data record developed and generated by EUMETSAT CM SAF, SMHI, which has also been produced from this FDR.

As a pre-condition, (FSC (NDSI) > 5%) was implemented based on Normalized Difference Snow Index (NDSI) calculations (Salomonson and Appel, 2006) before utilising the SCAmod algorithm (Metsämäki et al. 2015) for viewable snow (SCFV) and snow on ground (SCFG). A pre-classification was implemented to minimise erroneous results (solar zenith angle > 88°; cloud probability > 80%; water if percentage of pixel > 50; and permanent ice if percentage of pixel > 50). In addition, two thresholds were included to test whether a pixel potentially is snow free or snow covered (snow free if channel 1 > 0.12 or channel 4 > 283 K).

In addition, different post-processing steps were implemented to improve the quality of the snow products: for latitudes ±15° and elevations below 1000 m a.s.l. a test was added (channel 1 < 0.30 or channel 4 > 270 K) to remove erroneous data in the tropical regions. A second test considers channel 3b: (channel 1 – channel 3b ≥ 0.2 and channel 3b < 0.1) for the Southern Hemisphere to remove erroneous snow pixels, with an additional adaptation for the globe (channel 1 – channel 3b < 0.1 and channel 3b < 0.2). The erroneous snow pixels with (channel_3a(/b)/channel_1 > 1 and channel_3a(/b)/channel_1 > 1) or (channel_2/channel_1 < 0.999 and channel_4 > 274 K) was removed. Additionally, the pixels with high solar zenith angle (VZA > 75°) and NDSI < 0.5 was removed. Some other criteria were also added in post-processing steps.