Application of artificial neural network model for the identification the effect of municipal waste compost and biochar on phytoremediation of contaminated soils

Abstract

This research was carried out to assessing the potential of Bromus tomentellus for phytoremediation with biochar and municipal waste compost amendments to improving the clean-up efficiency of soils contaminated with chromium (Cr) and zinc (Zn). Soil amendment was added to contaminated soil in three levels (%0: Control; without organic fertilizer, biochar and compost 1%, biochar and compost 2%). It also determines the applicability of artificial neural network (ANN) in the modeling of the extraction process. The physiochemical properties of the contaminated soil, including pH, Electrical Conductivity (ECe), Cation Exchange Capacity (CEC) and Sodium Adsorption Ratio (SAR) were determined. After validation of the applied artificial neural network, the effect of municipal waste compost and biochar treatment on the absorption of heavy metals in different parts of the plant was investigated. Also, the range of adding amending factors to the soil through the neural network increased from 2% in experimental data to 5% in predicting data. The neural network was taught for heavy metals in soil and plant, so the amount of Correlation coefficient (R2) value in most cases was higher than 0.9 and close to 1 which means the Group Method of Data Handling (GMDH) and artificial neural network was usable for over-predicting data. The results indicated that by adding compost percentage, the absorption of Zn is also increased. The highest concentration of Zn (274.82 mg/kg) and Cr (26.66 mg/kg) was observed by adding 0.8% compost and 0.52% biochar, respectively. The maximum Cr concentration for compost (25.19 mg/kg) was detected by adding 1% compost.