[Quantifying direct N2O emissions from paddy fields during rice growing season in China: model and input data...

The models on direct N2O emissions from rice paddies under different water regimes developed by the authors were validated against field measurements in China reported in 2005-2007 and in other regions. In flooding rice paddies (F), N2O emission predicted by the model was consistent with previous reports in other regions. Under the water regime of flooding-midseason drainage-reflooding (F-D-F), the model developed in this study was comparable to that established by using worldwide database. The models also well fitted N2O emissions from rice paddies under the water regime of flooding-midseason drainage-reflooding-moisture but without waterlogging (F-D-F-M) in China. Consistency of rice production data derived from the database of this study with those reported in previous studies suggests that the model input data of rice production had high reliability. The input data showed that water management and nitrogen input regimes have greatly changed in rice paddies since the 1950s. During the 1950s-1970s, about 20%-25% of the rice paddy was continuous water logging, and 75%-80% under the water regime of F-D-F. Since the 1980s, about 12%-16%, 77% and 7%-12% of paddy fields were under the water regimes of F, F-D-F and F-D-F-M, respectively. Total N input during the rice growing season averaged 87.49 kg x hm(-2) in the 1950s and 224.64 kg x hm(-2) in the 1990s. Chemical N input during the rice growing season has increased from 37.4 kg x hm(-2) in the 1950s to 198.8 kg x hm(-2) in the 1990s, accounting for 43% and 88% of the seasonal total N inputs, respectively. Manure N input was applied at stable rate, ranging from 45.2 kg x hm(-2) to 48.2 kg x hm(-2) during the 1950s-1970s, but thereafter it decreased over time. The contribution of manure N to total N inputs has decreased from 52% in the 1950s to 9% in the 1990s. Crop residue N retained during the rice growing season has increased from 4.9 kg x hm(-2) in the 1950s to 6.3 kg x hm(-2) in the 1980s. A high spatial variation of nitrogen inputs during the rice growing season was found in the 1950s-1970s, while it was not pronounced in the 1980s-1990s. Overall, the results of this study suggest that the models could be used to quantify direct N2O emissions from rice paddies under various water regimes in China.