Author(s): D. Panagoulia; N. Maratos

Linked Author(s):

Keywords: Rtificial neural network; Linear least squares and multi-start simplex optimization; Conceptual modeling; Rainfall-runoff process; Mountainous catchment

Abstract: An algorithm using a combination of linear least squares and multi-start simplex optimization (LLSSIM), originally proposed by Hsu et al., 1995, is used in order to show the mechanism and parameters of three-layer feed forward ANN models and the potential of such models for simulating and forecasting the nonlinear hydrological behavior of mountainous catchments. The output ‘rain plus melt' from the snow accumulation and ablation model (SAA) of the US National Weather Service (US NWS) applied on a medium-sized mountainous catchment (the Mesochora catchment in Central Greece) was used as input to ANN model. The nonlinear ANN model approach is shown to provide a better representation of the rainfallrunoff relationship in medium and extreme conditions than the conceptual soil moisture accounting (SMA) model of the US NWS applied over the same catchment. Because the ANN approach presented here has not physically realistic components and parameters, it is by no means a substitute for conceptual catchment modeling. However, the ANN approach does provide a viable and effective alternative for developing input-output simulation and forecasting models in cases that do not require modeling of the internal dynamics of the catchment.

DOI:

Year: 2003