THE USE OF BACKPROPAGATION NEURAL NETWORKS FOR THE SIMULATION AND ANALYSES OF TIME-SERIES DATA IN SUBSURFACE DRAINAGE SYSTEMS

This study was undertaken to investigate the application of artificial neural networks (ANNs) in the simulation of subsurface drainage syste ms. Back-propagation ANNs were trained to imitate a conventional mathe matical model, DRAINMOD, in the simulation of water. table depths. For good representation of the dynamics of a soil system, the time lag pr ocedure was developed to feed the input values of previous time steps. The results show that the use of time lag procedures produced signifi cant impacts on the ANN performances. In this study, two methods are i ntroduced to analyze and compare the impact of various strategies of d ata input into the ANNs and DRAINMOD. In the model-response analysis, one input was varied with different impulses while the other inputs we re kept constant. The results showed that the optimal time-dependence period of the ANN inputs should be determined by the saturated hydraul ic conductivity and actual distance from the soil surface to the imper meable layer In the sensitivity analysis, each processing element in t he input layer of an ANN with the lag procedure was disabled respectiv ely. When the processing elements corresponding to the inputs of rainf all and previous water table were disabled, the r(2) values of linear regression could decrease to less than 0.1. The results showed that th ese inputs were more important to the ANNs. These methods can be used to evaluate the performances of simulation models for time-series data and real-time control, particularly when the real situation is not av ailable.