Artificial neural networks (ANN) model was used to estimate N2O gas flux emitted from rice paddies with several water regime treatments. The purpose of this study were to identify the relationship of micro-environment with different water regimes towards N2O gas flux, to predict the amount of N2O gas flux, and to validate neural network models. Field experiment was conducted in the field laboratory of Civil and Environmental Engineering-IPB from February to August 2016. Land of paddy was treated with three water regime of continuous flooding water regime, wet water regime, dry water regime. ANN model with back propagation algorithm consisted of input layer with six nodes of micro- environment parameter and N2O gas was used as the model output. The total N2O flux for continuous flooding water regime, wet and dry regime were -25.95 mg/m2/season, 17.32 mg/m2/season, and 21.16 mg/m2/season. ANN each water regime was obtained the coefficient of determination (R2) of was 1, so ANN model was acceptable and could be used to predict N2O flux. Key words: artificial neural networks, micro environment, N2O gas flux, paddy field
Copyrights © 2019