Abstract: Methane emission fluxes from late rice fields under different fertilization treatments were studied, using the static chamber-gas chromatograph techniques. Results show that compared to Treatment 1 (T1,no fertilization), all the other fertilization treatments increased methane emission. The average methane flux from Treatment 5 (T5,straw incorporation) and Treatment 4 (T4,increased nitrogen and phosphate) was 9.96 and 7.90 mg·m^-2·h^-1, respectively. Treatment 2 (T2,balanced fertilization) and Treatment 3 (T3,decreased nitrogen and phosphate) were just approximately 20% more than Treatment 1. In terms of accumulated seasonal methane emission, the treatments followed an decreasing sequence: T5>T4>T2>T3>T1, but no significant difference (P>0.05) was observed between T1,T2,T3 and T4. According to the analysis of related environmental factors, methane emission fluxes had an exponential function relationship with soil temperature at 5 cm depth, and had a positive correlation with thickness of the water layer (P<0.05). T2 was a fertilization model recommended for the local region after comprehensive consideration of the greenhouse effect and rice yield. Considering global warming potentials of CH₄ and rice yield, T2 was a recommended fertilization method under the local condition. Total fertilizer application of N, P₂O₅ and K₂O was 180,90 and 135 kg·hm^-2, respectively. Ammonium bicarbonate, with a split 70% of designed N mixed with total P and K₂O (calcium superphosphate and potassium chloride) was broadcasted on the field before transplanting as basal fertilizer,and urea was topdressed with 30% of total N at the tillering stage.

Key words: late rice field, fertilization, methane, global warming potential, yield