Mountain regions are considered to be the natural "water towers" of the world due to their importance as sources of many rivers. Reliable tools to estimate the availability and variability of streamflows in such regions are still rare. In this context, the present Thesis proposes to extend an existing Flow Duration Curve (FDC) modelling framework to Alpine environment. FDCs can be obtained empirically or based on models. Process-based FDC models have the advantage of incorporating hydrological process knowledge and thereby allowing the prediction of FDCs under changing conditions. This Thesis studies a simple process-based model that describes daily streamflow distributions as the result of subsurface flow pulses triggered by stochastic rainfall and censored by the soil moisture dynamics. The resulting streamflow distribution is characterized by only a few parameters: the mean rainfall depth and the frequency of events that produce discharge and recession parameters. The objective of this Thesis is the extension of the existing framework (Botter et al, 2007) for pluvial streamflow regimes to Alpine environments where the accumulation of water in the form of snow and ice influences the streamflow regime. The selected study region is Switzerland, an Alpine country with a wide range of hydro-climatologic conditions. The key of the extension of the model framework is a seasonal approach, i.e., a model set up for each of the up to three distinct seasons encountered in Alpine environments: i) pluvial season, ii) accumulation season, and iii) melting season. The pluvial season can be modelled by the original model framework, but required the definition of more robust parameter estimation methods in the context of this Thesis, particularly for the linear and nonlinear recession parameters. The extension to the melting season is newly developed in this Thesis, incorporating the earlier extension to the snow accumulation season by Schaefli et al, 2013. A key result of all completed parameter estimation tests for pluvial regimes is the very good performance of an inverse approach based on maximum likelihood estimation (MLE). MLE shows outstanding results even for short series of observations and can be retained as the recommended method to be used for the model framework studied in this Thesis. The extension to the ablation season is achieved by the incorporation of the melting contribution as equivalent precipitation and an ensuing increase of the streamflow producing frequency as compared to the one resulting from rainfall input. The amount of equivalent precipitation is calculated based on the snow accumulation from the existing model extension to winter low flows, combined with a process-based definition of seasons rather than calendar dates. A detailed analysis for all seasons for 10 selected case studies shows that the new seasonal approach yields good results for Alpine streamflow distributions, including for glacier-influenced catchments. The improved parameter estimation methods developed in this Thesis establish a new reference approach for regionalization, opening new perspectives for flow duration curve estimation in ungauged catchments. Other promising results are the consistency of estimated model parameters with underlying physical processes and namely the observed correlation between model parameters and mean catchment elevation. This will allow the study of land use and climate changes in future model applications.