Author(s): Kun Guo; Yoon-Seok Timothy Hong; Rao Bhamidimarri; Daqing Chen
Linked Author(s):
Keywords: Hybrid modelling; Rainfall-recharge system; Soil moisture balance; Neural Network
Abstract: The estimation of groundwater recharge is of critical importance to the management of groundwater resources. One of the most common rainfall-recharge models is soil moisture balance model. However, its performance sometimes is not good enough because of our limited knowledge about the mechanisms of the rainfall-recharge system. To enhance the performance of the soil moisture model, a hybrid model is proposed, which uses neural network to correct output error of the soil moisture balance model. Data is collected from four sites in New Zealand and an application of the proposed hybrid model is carried out in these sites. Simulation results show that the performance of the proposed hybrid model is generally better than that of the original soil moisture balance model and multilayer perceptron. Due to the good performance of the proposed hybrid model, it is not only suitable for hydrological modeling, but also has potential to be applied in a variety of complex modeling fields.
Year: 2013