Assessing Ambiguity in Physically-based Hydrological Models: Stochastic Approach with Application to the Ardeche Catchment (Southern France)

摘要

Within the framework of physically-based hydrological modelling, the estimation of the parameter values and the simulation reliability are often influenced by measurement uncertainties and lack of information. These problems affect the values of the input variables (such as topography and precipitation), the model parameters (such as the hydraulic conductivity) and the 'target' variables (such as the river discharge). Hence, even when the processes are correctly simulated, the solution of the inverse problem of parameter values estimation is often non-unique. A stochastic technique is suggested for the estimation of different sets of model parameters to be considered as ambiguous, i.e. giving simulations that cannot be statistically considered as significantly different.