Prediction Of Hydrological Models’ Uncertainty By A Committee Of Machine Learning-Models

This study presents an approach to combine uncertainties of the hydrological model outputs predicted from a number of machine learning models. The machine learning based uncertainty prediction approach is very useful for estimation of hydrological models' uncertainty in particular hydro-metrological situation in real-time application [1]. In this approach the hydrological model realizations from Monte Carlo simulations are used to build different machine learning uncertainty models to predict uncertainty (quantiles of pdf) of the a deterministic output from hydrological model . Uncertainty models are trained using antecedent precipitation and streamflows as inputs. The trained models are then employed to predict the model output uncertainty which is specific for the new input data. We used three machine learning models namely artificial neural networks, model tree, locally weighted regression to predict output uncertainties. These three models produce similar verification results, which can be improved by merging their outputs dynamically. We propose an approach to form a committee of the three models to combine their outputs. The approach is applied to estimate uncertainty of streamflows simulation from a conceptual hydrological model in the Brue catchment in UK and the Bagmati catchment in Nepal. The verification results show that merged output is better than an individual model output. [1] D. L. Shrestha, N. Kayastha, and D. P. Solomatine, and R. Price. Encapsulation of parameteric uncertainty statistics by various predictive machine learning models: MLUE method, Journal of Hydroinformatic, in press, 2013.

Combined Vertical And Lateral Channel Evolution Numerical Modeling

Vertical stream bed erosion has been studied routinely and its modeling is getting widespread acceptance. The same cannot be said with lateral stream bank erosion since its measurement or numerical modeling is very challenging. Bank erosion, however, can be important to channel morphology. It may contribute significantly to the overall sediment budget of a stream, is a leading cause of channel migration, and is the cause of major channel maintenance. However, combined vertical and lateral channel evolution is seldom addressed. In this study, a new geofluival numerical model is developed to simulate combined vertical and lateral channel evolution. Vertical erosion is predicted with a 2D depth-averaged model SRH-2D, while lateral erosion is simulated with a linear retreat bank erosion model developed in this study. SRH-2D and the bank erosion model are coupled together both spatially and temporally through a common mesh and the same time advancement. The new geofluvial model is first tested and verified using laboratory meander channels; good agreement are obtained between predicted bank retreat and measured data. The model is then applied to a 16-kilometer reach of Chosui River, Taiwan. Vertical and lateral channel evolution during a three-year period (2004 to 2007) is simulated and results are compared with the field data. It is shown that the geofluvial model correctly captures all major erosion and deposition patterns. The new model is shown to be useful for identifying potential erosion sites and providing information for river maintenance planning.