Validation of revised DNDC model for methane emissions from irrigated rice fields in Thailand and sensitivity analysis of key factors

The original Denitrification-Decomposition ( DNDC) model and a revised version were tested against data from field observations of methane (CH4) emissions from irrigated rice fields in Thailand. The revised DNDC model, which was modified for use in Japanese rice fields by revising the crop growth and soil biogeochemical submodels, yielded better simulation results than the original model. In most cases, daily CH4 fluxes predicted by the revised DNDC model agreed well with observations. Seasonal CH4 emissions simulated by the revised model showed significantly higher correlation with observations than those obtained with the original model. Errors in the simulation appear to have resulted from uncertainties in both the input parameters and the model descriptions. Sensitivity analysis revealed that the revised DNDC model is highly sensitive to the concentration of reducible soil Fe, the rate of rice straw incorporation, and rice root biomass. Therefore, uncertainties in these factors may strongly affect the prediction of CH4 emissions. These results suggest that for reliable prediction of CH4 emissions from Thai rice fields, further work is needed to improve the estimates of reducible soil Fe, to quantify the rate of straw incorporation, and to parameterize the crop submodel for the dominant rice varieties grown in Thailand.