<![CDATA[The paper proposes the artificial neural networks (ANN) model to predict methane (CH4) and Nitrous Oxide (N2O) emissions under different irrigation system based on easily measurable environmental biophysics parameters such as soil moisture, soil temperature and soil electrical conductivity. To verify the model, two experiments were conducted in the pot experiments in two different locations. The first location was in the greenhouse of Meiji University, Kanagawa Prefecture, Japan from 4 June to 21 September 2012, and the second location was in water resources engineering laboratory, Department of Civil and Environmental Engineering-IPB from 2 July to 10 October 2014. In each location, there were three different irrigation systems adopted with the System of Rice Intensification (SRI) principles. We called the experiment as SRI Basah (SRI B1 and SRI B2 for the first and second locations, respectively), SRI Sedang (SRI S1 dan SRI S2) dan SRI Kering (SRI K1 dan SRI K2). Each treatment has different water level during growth stages. As the results, the developed ANN model can predict CH4 and N2O emissions accurately with determination coefficients of 0.93 and 0.70 for CH4 and N2O prediction, respectively. From the model, characteristics of those greenhouse gas emissions can be well identified. For the mitigation strategy, SRI B1 and SRI B2 treatments in which the water level was kept at nearly soil surface are the best strategy with highest yield production and lowest GHG emission.]]>
Copyrights © 2015
