Summary Summary

Distributedhydrological modelling is useful to improve our understanding of the mechanics ofnatural processes in a watershedand their interaction with human activity. It can support the estimation ofwateravailability, and the assessment ofthe impacts of climate and land use change or other activities within a watershed. This can support decision-making about measures to improve flood protection, water quality, aquatic ecosystems, and potential for recreational activities in the watershed. As all hydrologic modellingresults are subject to uncertaintydue to measurement errors in input and response and error in model strueture, the reliability of modellingresults must be assessed by estimating their uncertainty. In the last two decades, many uncertainty analysis techniques were developedand applied in the field of hydrology. Most of the uncertainty analysis (UA) methodologiesfocus only on parameter uncertainty and other sources of uncertainties are not or only partially represented,and the advantage and disadvantage ofdifferent UA techniques are not comparatively investigated. The aim of this study is to developa UA methodology which describes the effect ofboth parameter uncertainty and other sources of uncertainty and combines prior knowledge about parameter values with empirical evidence from the catchment to reduce predictio