We have applied three models, a neural network, a conceptual model and a combination of these two a hybrid model, to model the backwater effect of ice in a river. The neural network is a black-box model. It is based mainly on observed data and it lacks the expert knowledge of the system. The conceptual model is based on a physical description of the system. The data is used in optimizing the free parameters of the description. In the hybrid model, the neural network is modified so that the physical description of the conceptual model can be coded into the structure of the network. In the beginning of fitting, the hybrid network already performs as well as the conceptual model. During fitting also the structure of the physical description is optimized, not only the parameters of the description. The three models are rather different in form but in the modeling results there are only slight differences. Mean error of the models in ice-correction is 13-15 m3/s at an observation station where the mean backwater effect of the ice is 100 m3/s. The aim of this work is to develop a model for real time estimation of corrected discharge, which is used in error correction of a discharge forecast model. For this purpose the error of the best model is acceptable.
Neural Networks in the Ice-Correction of Discharge Observations: Paper presented at the Nordic Hydrological Conference (Akureyri, Iceland–August 1996)
Markus Huttunen, Bertel Vehviläinen, Esko Ukkonen; Neural Networks in the Ice-Correction of Discharge Observations: Paper presented at the Nordic Hydrological Conference (Akureyri, Iceland–August 1996). Hydrology Research 1 August 1997; 28 (4-5): 283–296. doi: https://doi.org/10.2166/nh.1998.21
Download citation file: