Estimating Greenhouse Gas Emissions from Irrigated Paddy Fields in Indonesia under Various Water Managements

1st International Conference on Mechanical Electronic and Biosystem Engineering

Arif, Chusnul^1 ; Indra Setiawan, Budi^1 ; Iswati Hasanah, Nur Aini^2 ; Mizoguchi, Masaru^3

Department of Civil and Environmental Engineering, IPB University, Indonesia^1

Department of Environmental Engineering, Islamic University of Indonesia (UII), Yogyakarta

55584, Indonesia^2

Department of Global Agricultural Sciences, University of Tokyo, Tokyo

113-8657, Japan^3

关键词: Artificial neural network models;    Estimating method;    Field experiment;    Measurable parameters;    Measurements of;    Soil electrical conductivity;    Soil redox potential;    Soil temperature;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/557/1/012034/pdf
DOI  :  10.1088/1757-899X/557/1/012034

来源: IOP

【 摘 要 】

Methane (CH4) and nitrous oxide (N2O) are two main greenhouse gasses emitted from paddy irrigated paddy fields. Their fluxes are commonly affected by water managements in the fields. However, the main problem in the study of greenhouse gas emissions in paddy fields is the instrumentation for measuring emissions. Measurements of greenhouse gas emissions are costly and complicated. The current study proposes estimating method to quantify greenhouse gas emissions by an artificial neural network (ANN) model. They are estimated based on easily measurable parameters such as soil moisture, soil temperature, soil electrical conductivity (EC), soil redox potential (Eh) and soil pH. The model was verified based on field experiments that were conducted in Bogor, West Java, Indonesia during 26 March - 24 June 2015. Here, three regimes of water management, i.e. continuous flooded (FL), moderate (MR) and dry (DR) regimes, were performed in the field. The DR regime released the lowest total greenhouse gas emissions; however, it reduced grain yield by 58% and 12% compared to the FL and MR regimes respectively. The developed model showed high accuracies for both greenhouse gasses estimation where the coefficients of determination (R2) values were 0.84 and 0.76 for CH4 and N2O prediction respectively.