The scientists developed a method to estimate how much snow is melting from a snowpack, which is important for safe train operation during the melting period. They created four models to estimate the snowmelt volume using data from a station near Oshirakawa. The models had an error of around 10% when compared to actual measurements, with the physical model having an error of 0.66 mm/h and the statistical model having an error of 0.173 mm/10 min. This means the models can help predict how much snow is melting, which is crucial for preventing accidents on train tracks.