Search

This paper presents a systematic investigation into modelling capacities of three conventional data-driven modelling techniques, namely, wavelet-based artificial neural network (WANN), support vector regression (SVR) and deep belief network (DBN) for multi-step ahead stream flow forecasting. To evaluate the effectiveness of these modelling techniques, hydro-meteorological hourly datasets from three case-study rivers located in the UK have been used. A heuristic performance analysis of the modelling schemes has been conducted by systematically analysing the key statistics that measure magnitude, scatter and density of model errors. Finally, for each of the modelling techniques, the performance deterioration rate in time was estimated. The results show that the SVR model can forecast quite accurately up to one to two hours ahead but its performance deteriorates gradually from three hours onwards. Further it has been found that the WANN model performs better when the overall non-linearity of the system increases, whereas the DBN model appeared to show consistently poor predictive capabilities when compared to the other models presented herein. The authors conclude by stating that, for any selected model, it is possible to use an identical model structure for up to two steps ahead forecasting. Models need to be re-configured beyond that limit. [ABSTRACT FROM AUTHOR]

Fluvial flows and tidal flows are governing on tidal rivers while fluvial flows are governing on non-tidal rivers alone. In addition to analytical and numerical models, stochastic methods must be considered to determine water surface elevation at the tidal limit of tidal rivers because hydraulic routing of this reach is very complex. In the present research, the neural network method was considered for determination of water surface elevation at the tidal limit of tidal rivers. The artificial neural network (ANN) method was trained using results from a suitable numerical model. The results of the ANN method were compared with results for the governing regression relation for water surface elevation at the tidal limit of tidal rivers. The ANN method was applied to the Karun river in Iran and the River Severn in the UK. The ANN method was found to produce water surface elevations for different combined return periods in the Karun river and the River Severn. [ABSTRACT FROM AUTHOR]