A dynamic model of snowpack density, depth and mass content and its validation with Snotel hourly data

The knowledge and the prediction of snowpack mass content (in the form of snow water equivalent) represent a crucial information for many engineering purposes as, for example, hydrological applications, as well as social, territorial and civil engineering aims. Furthermore, snowpack characteristics are strongly sensitive to climate change, so that a precise and reliable modeling has to assist policy-makers in predicting and evaluating future scenarios. In addition, snow density appears at the moment to be one of the most problematic snowpack state variable. As a consequence, this dissertation deals with an attempt to characterize snow density dynamics throughout the existence period of a generic snowpack. Besides, a new model for the coupled dynamics of snow density and snow depth at the hourly scale is proposed, in order to provide a precise time-variant description of snowpack mass content. Thirdly, a multi-year scale model validation is presented, adopting as input variables the measured data series of an instrumental network which covers the Western United States (SNOTEL). As a result, the importance of separating mechanic forcings from the hydraulic ones is stressed, in order to correctly predict the coupled evolutions of the snowpack structure and of its liquid content. Secondly, a physical-based, innovative and computationally easy model has been provided. In addition, 49 SNOTEL sites have been chosen to validate it. Since snow depth hourly data were found to be really noisy, a three stage process has been designed to edit snow depth data and to reconstruct more plausible precipitation input data series. As an innovative issue, a refined procedure has been designed to eliminate temperature-based snow depth oscillations. As a consequence, simulations turned out to be rather satisfying and site-adaptive.