ABSTRACT
The use of tires in the constructed wetlands gives environmentally suitable solutions for disposal problems, which end up accumulating in rivers and public designations. Others burned tires, releasing gases into the atmosphere. For gravel media, the mining, transportation, and handling of this conventional media are expensive and the clogging problem may be appeared after short time from operation. However, the gravel media is suitable for cultivation and root penetration. Rubber, gravel, and plastic media; biochemical oxygen demand and chemical oxygen demand and fecal coliforms and Cadmium gave non-significant difference. The remaining parameters showed a significant difference. Both artificial neural network and statistical linear and nonlinear equations through the stochastic package for social science (SPSS) package managed to mimic the horizontal subsurface flow wetlands for all used parameters with an acceptable accuracy. The artificial neural network models represent the experimental data in calibration and validation processes more accurately than the SPSS models.
