Modelling of Greenhouse Gas Emissions in Irrigated Lowland Rice by the Alternate Wetting-Drying (AWD) System

Abstract

Greenhouse gas (GHG) emissions have become a major global environmental issue due to their significant contribution to climate change. In irrigated rice fields, methane (CH 4 ) and nitrous oxide (N 2 O) are the dominant greenhouse gases influenced by water management practices. The Alternate Wetting-Drying (AWD) irrigation system has been developed as an alternative to continuous flooding to reduce GHG emissions while maintaining rice productivity. Therefore, this study aimed to develop a GHG emission model under AWD systems using biophysical environmental parameters and an Artificial Neural Network (ANN), as well as to compare emission characteristics under different irrigation regimes. Field observations were conducted under three irrigation regimes, namely continuous flooding (FL), wetting (WT), and drying (DR) representing AWD conditions. CH 4 and N 2 O emissions were measured weekly using standard methods, while soil temperature, soil water content, electrical conductivity (EC), pH, and redox potential (Eh) were used as model input variables. The ANN model demonstrated excellent performance in predicting CH 4 emissions, with a coefficient of determination (R 2 ) of 0.92, while N 2 O prediction showed good performance with an R 2 value of 0.82. These results indicate that ANN is effective for modelling greenhouse gas emissions in irrigated rice fields, particularly for CH 4 emissions.