Data-driven approaches for estimating uncertainty in rainfall-runoff modelling

This paper presents application of the method for assessment of predictive uncertainty of a rainfall-runoff model using data-driven modelling techniques. Historical model errors which are mismatch between observed and simulated values of runoff by rainfall-runoff models are assumed to be indicators of total model uncertainty. Uncertainty is represented by the model prediction interval. Different data-driven techniques used to determine the interval are compared. The method is tested to estimate uncertainty of runoff simulations by conceptual rainfall-runoff model with application to Brue catchment in the United Kingdom; it is also compared with other uncertainty estimation methods.