Abstract
Understanding the dynamics of snowmelt is essential for hydrological forecasting, natural hazard assessment, and water resource management in both mountainous headwater catchments and downstream lowland systems. In recent years, multiple strategies have emerged to assess melt onset and dynamics, including in situ measurements, physically based snowpack models, and microwave and thermal satellite data. However, few studies have directly compared these approaches to evaluate their consistency in detecting snowmelt phases. This study presents a multitemporal analysis (2016–2021) integrating three different snow detecting methods over two Alpine sites in the Aosta Valley (Western European Alps): Torgnon (2160 m a.s.l., subAlpine) and Cime Bianche (3100 m a.s.l., Alpine). The dataset includes: (i) ground meteorological measurements from Automatic Weather Stations (AWS), used to force the physically based SNOWPACK model and obtain key snowpack variables such as snow temperature, liquid water content (LWC), and snow water equivalent (SWE); (ii) optical and thermal infrared measurements used to retrieve the Apparent Thermal Inertia of snow (APs), which is sensitive to variations in surface temperature and albedo; and (iii) C-band Synthetic Aperture Radar (SAR) backscatter time series from Sentinel-1 (orbits 66, 88, and 139), sensitive to changes in liquid water content and snow structure. We focused on identifying the three characteristic phases of snowmelt—warming, ripening, and output—defined by specific thresholds. The onset, end, and duration of each phase were extracted independently from the three datasets and compared. Findings show that, although based on different physical principles, radar, thermal, and model-based approaches provide consistent information on snowmelt timing when systematically compared. In particular, the comparison between Sentinel-1 SAR backscatter and APs highlights their ability to detect the onset and progression of melt phases in a coherent way. Results show that melt progression is sitedependent: at the sub-Alpine site, melt phases begin earlier and are shorter, whereas the Alpine site exhibits delayed and prolonged transitions due to colder conditions and wind-driven snow redistribution. Future satellite missions such as ESA’s LSTM and NASA’s SBG-TIR will significantly enhance thermal monitoring of the snow surface, while current hyperspectral missions like PRISMA and EnMAP enable the retrieval of key snow properties (e.g. light-absorbing impurities and liquid water content). These developments open new perspectives for remotely sensing snowmelt phases across both mid-latitude and polar regions.