Modeling biodegradation and kinetics of glyphosate by artificial neural network

Abstract

An artificial neural network (ANN) model was developed to simulate the biodegradation of herbicide glyphosate [2-(Phosphonomethylamino) acetic acid] in a solution with varying parameters pH, inoculum size and initial glyphosate concentration. The predictive ability of ANN model was also compared with Monod model. The result showed that ANN model was able to accurately predict the experimental results. A low ratio of self-inhibition and half saturation constants of Haldane equations (< 8) exhibited the inhibitory effect of glyphosate on bacteria growth. The value of K_i/K_s increased when the mixed inoculum size was increased from 10⁴ to 10⁶ bacteria/mL. It was found that the percentage of glyphosate degradation reached a maximum value of 99% at an optimum pH 6-7 while for pH values higher than 9 or lower than 4, no degradation was observed.

Keywords:

• Glyphosate
• biodegradation
• Monod model
• Haldane model
• artificial neural network (ANN)

Acknowledgment

The authors gratefully acknowledge Crop Protection (M) Sdn. Bhd for providing the glyphosate throughout this research study.