Objective To investigate the seasonal variation characteristics of methane(CH₄) emission flux in paddy fields and analyze the main influencing factors and their quantitative contributions.

Method The seasonal variation characteristics of CH₄ emission flux in paddy fields were analyzed using the in-situ observation data from the eddy covariance system of farmland ecosystems at Jurong Ecological Experimental Station, Jiangsu Province in 2016. The influencing factors and their quantitative contributions of the seasonal variation of CH₄ emission flux in paddy fields were analyzed based on correlation analysis and path analysis methods.

Result The CH₄ emission flux in paddy fields increased rapidly from the average value of 129.82 mg/m²·d in mid-June to the peak value in late July and early August (1 191.78 mg/m²·d), representing a 9.18-fold increase, and then gradually decreased to a low value at the end of the rice growing season in early December (4.35 mg/m²·d), and experienced several significant fluctuations during the rice growing season. The correlation between CH₄ flux and 11 relevant environmental variables (TA, TS, PPFD, VPD, μ^*, AP, RH, LAI, GPP, LE, WTD) differed significantly between the pre-and post-panicle initiation periods of the rice-growing season. During the pre-panicle initiation of the rice-growing season, the direct path coefficients of TS, TA, and GPP for CH₄ fluxes were 0.72, -0.45 and 0.76, respectively. TA indirectly affects the seasonal variation of CH₄ emission flux through TS and VPD; LAI indirectly influences the seasonal variation of CH₄ emission flux through GPP and LE, with indirect path coefficients of 0.94 and 0.69, respectively. TS and LAI had a strong influence on the seasonal variation of CH₄ fluxes during post-panicle initiation of the rice-growing season, and the direct path coefficients were 0.73 and -0.60, respectively. LAI can also influence the seasonal variation of CH₄ emission flux through GPP and LE, with indirect path coefficients of -0.45 and -0.51 respectively. The indirect path coefficient of TA influencing the seasonal variation of CH₄ emission flux through TS is 0.93.

Conclusion The CH₄ emission flux in paddy field ecosystems had obvious seasonal variations. In the pre-panicle initiation of the rice-growing season, TS, TA, and GPP were the main influencing factors of the seasonal variation of CH₄ flux, however, during the post-panicle initiation of the rice-growing season, TS and LAI were the main influencing factors of the seasonal variation of CH₄ flux. The relationship between influencing factors and daily CH₄ fluxes was well fitted by simple linear or exponential functions and with good goodness of fit.