Characterization of layers of snowpack with highly resolved micro-cone penetration tests leads to detailed fluctuating signals. We used advanced stochastic analysis to differentiate snow types by interpreting the signals as a mixture of continuous and discontinuous random fluctuations. These two types of fluctuation seem to correspond to different mechanisms of drag force generation during the experiments. The proposed methodology provides new insights into the characterization of snow layers.