*2.4. Orthorectification*

The orthorectification module was based on the geometrical correction of the perspective view. This step was implemented following [32]. The available digital elevation model [33], with a 5 m spatial resolution and 1 m vertical resolution, provided about 300 topographic points that were projected on the camera view (Figure 1c). The effectiveness of the correction was estimated considering eight ground control points.

#### *2.5. Satellite Snow Products*

Several satellite products are available for the remote sensing of the cryosphere and for this study we considered products obtained by optical sensors, characterized by different spatial resolutions: high (below 100 m); intermediate (below 1 km); and low (higher than 1 km). The integration between those products and ground-based imagery will be tested, in order to improve the dataset concerning the snow cover over a decade.

#### 2.5.1. Optical Remote Sensing with High Spatial Resolution

The available remotely sensed snow products with a higher spatial resolution (below 100 m) were limited to Landsat missions, considering the studied time range (2004–2013). The selected sensors included Landsat satellites from 5 to 8, taking some differences into account in terms of band spectral ranges. All these data are now processed and available in the Swiss Data Cube [34]. The Landsat satellites are characterized by a spatial resolution of 30 m and a revisit time of 16 days. The considered data were geometrically and atmospherically corrected (Level 2A) using the Second Simulation of the Satellite Signal in the Solar Spectrum (6S) algorithm available in the Atmospheric and Radiometric Correction of Satellite Imagery (ARCSI) software [35]. The final estimation of NDSI was possible considering Eq.1 and the first short-wave infrared band of Landsat sensors. The wavelength ranges are specific for each sensor and they correspond to 520–600 nm and 1550–1755 nm for missions 5 and 7, and 533–590 nm and 1566–1651 nm for mission 8 [36].

#### 2.5.2. Optical Remote Sensing with Intermediate Spatial Resolution

The highest time resolution available for optical remote sensing at our latitudes is provided, in the framework of the Earth Observing System (EOS) flagship, by NASA's satellites Terra and Aqua. Both platforms are equipped with the Moderate Resolution Imaging Spectroradiometer (MODIS) and they provide the coverage of the Earth two times daily (Terra in the local morning and Aqua in the local afternoon). The instrument is characterized by 36 bands with a spatial resolution of 250 m in the visible range and 500 m in the short-wave infrared. The NASA's data chain provides the retrieval of NDSI at the ground, and we considered the MOD10A1\_006 and the MYD10A1\_006 products, for Terra and Aqua respectively, obtained using the National Snow and Ice Data Center services [11]. The NDSI values, calculated using the MODIS bands 4 (545–565 nm) and 6 (1628–1652 nm) in Equation1, were obtained in absence of clouds 2314 times over 6556 overpasses within the studied period.

#### 2.5.3. Optical Remote Sensing with Low Spatial Resolution

The daily estimation of the snow cover extent is being provided, over the considered period, by the European Space Agency as a component of the Data User Element. The GlobSnow Snow Extent (SE) product covers the Northern Hemisphere and it is going to be extended to the Sentinel missions. The GlobSnow SE processing system applies optical measurements in the visual and in the thermal part of the electromagnetic spectrum acquired by the ERS-2 sensor ATSR-2 and the Envisat sensor AATSR. The first step of the data chain is based on a cloud-cover retrieval algorithm (SCDA) where clouds, as well as large water bodies (oceans, lakes and rivers) and glaciers, are masked out. This algorithm is based on the brightness - temperature difference between 11 and 3.7 μm and on a set of additional rules, useful for certain sky conditions. Furthermore, the snow cover information is retrieved for not-vegetated areas by the NLR algorithm [37] where the band 2 (670 nm) is considered. This step is based on a semi-empirical reflectance model, where reflectance from a target is expressed as a function of the snow fraction. The Fractional Snow Cover can then be derived from the observed reflectance based on the given reflectance constants and the transmissivity values. The product is provided daily with a spatial resolution of 1 km and the data are available using the GlobSnow service [13].
