Application of Machine Learning Method to Predict Parameter Uncertainty

This chapter presents the application MLUE (machine learning method to predict parameter uncertainty) method to rainfall-runoff models. We use three machines learning methods, namely artificial neural networks, model trees; and locally weighted regression to predict several uncertainty descriptors of the rainfall-runoff model outputs. The generalised likelihood uncertainty estimation (GLUE) method is first used to assess the parameter uncertainty of the model and then machine learning methods are used to predict the uncertainty estimated by the GLUE method. The Hydrologiska Byråns Vattenbalansavdelning (HBV) conceptual rainfall-runoff model of the Brue catchment in the United Kingdom is used for the application of the MLUE method.