Gh.Ali Dezvareh - Ph.D. Candidate, Civil and Environmental Dept. of Eng. Faculty, KN Toosi University of Technology , Tehran, Iran
Azim Shirdeli - Assistant Professor, Water Engineering Dept., Zanjan University, Zanjan, Iran
Majid Ehteshami - Assistant Professor, Civil and Environmental Dept. of Eng. Faculty , KN Toosi University of Technology, Tehran, Iran

This study illustrates and compares the application of artificial neural networks (ANN) and CMLS Model for prediction of pesticide movement in Rey’s groundwater. A computer simulation investigate the ground water pollution potential of Rey area. The CMLS was able to simulate the movement of pesticides in multi layered soils. All factors including hydrogeology such as rainfall, temperature, and soil properties, and chemical characteristics such as partition coefficient and half-life were used in simulation process. With comparing the results of simulation and the groundwater depth the probability of pollution in different soil series, was simulated. Ten soil series were chosen with the probability of pesticides contaminations. In all soil series Oxydemrton methyl, Thimeton, Heptenphos azocyclotin, Bromopropilat, Phenpropatrin dinocab, Perimicarb, Trichlorfon, Captan, Benomyl, Etrimiphos, Carbendazim and Diazinon were used as pesticides. Diazinon was selected as one of the pesticides which could cause groundwater contaminant. Diazinon was detected through water sampling and GC analyses in Bagherabad soil series. The results showed the simulation of the CMLS model was reliable. ANN analysis also was performed to measure the relative importance of one input parameter over the others in pesticide groundwater occurrence. The prediction efficiencies such as Ef, RMSE, and R with a network of a four-layer BP was studied. The time of sample collection TSC; season, the groundwater depth, and pesticide travel times (PTT) were more sensitive parameters in the prediction of groundwater contamination.

ANN, Bagher Abad, CMLS, Groundwater, Pollution, Pesticides