Abstract: [Objective]This paper tries to reveal the impacting patterns and mechanisms of draining for forestation that affect carbon source or sink of temperate marsh wetlands at ecosystem level, in order to provide a scientific basis for carbon sink of wetlands management.[Method]The carbon (CO₂ and CH₄) emission fluxes, vegetation net primary productivity and carbon sequestration were studied by static chamber gas chromatography method and Multi N/C 3100,HT 1300 Solids Module (Analytik Jena AG, Germany) and the relative growth equation method from natural marsh wetland and Larix gmelinii plantation in the marsh wetlands under different draining treatments (10 years old, 30 years old) in Xiaoxing'an mountains of China, so as to evaluate the effects of wetland draining for forestation on carbon source or sink of temperate marshes which based on net ecosystem carbon balance.[Result]Draining for forestation changed the seasonal variation trend of CH₄ emission fluxes from natural marsh wetlands which turned from approximate unimodal type into alternative emission and uptake. And it also changed the function of CH₄ source or sink,from strong CH₄ emission source,mean annual fluxes were 1.780 mg·m^-2h^-1, to weak CH₄ uptake sink (-0.006 mg·m^-2h^-1)); Draining for forestation had no significant influence on the mean annual fluxes (168.07-220.43 mg·m^-2h^-1) of CO₂ emission from natural marsh wetlands; Compared with the natural marsh wetlands, mean annual fluxes of CO₂ emission decreased by 12.8%(P>0.05) from 10 years old Larix gmelinii plantation and increased by 14.3% (P>0.05) from 30 years old Larix gmelinii plantation; Draining for forestation changed the main controlling factors of CH₄ and CO₂ emission from natural marsh wetlands. CH₄ emission fluxes from natural marsh wetlands turned from significant positive correlation with 30-40 cm soil temperature, into that they were not related with 30-40 cm soil temperature after draining for forestation 10 or 30 years; the main controlling factors of CO₂ emission fluxes from natural marsh wetlands were the air temperature and the 0-30 cm soil temperature, but the drainage ridge of 10 or 30 years old plantation turned into air temperature and 0-40 cm soil temperature, while the main controlling factors of the drainage ditches from 30 years old plantation were significantly correlated only with air temperature and 0 cm soil temperature, drainage ditches of 10 years old plantation were not related with air temperature and soil temperature; The net vegetation primary productivity (15.44 t·hm^-2a^-1) and annual net carbon sequestration (6.74 t·hm^-2a^-1) of natural marsh wetlands were significantly reduced by 31.9% and 30.6% (P<0.05) after the draining for forestation for 10 years (10.51 and 4.68 t·hm^-2a^-1), but there were no significant difference between larch plantation and natural marsh wetlands (-6.7% and -5.2%, P>0.05) after the draining for forestation for 30 years (14.40 and 6.39 t·hm^-2a^-1); The carbon sink of natural marsh wetlands (2.08 t·hm^-2a^-1) significantly decreased by 65.4% (P<0.05) after the draining for forestation for 10 years (0.72 t·hm^-2a^-1), yet there was no significant gap between the larch plantation (1.20 t·hm^-2a^-1) and natural marsh wetlands (-42.3%, P>0.05) after the draining for forestation for 30 years.[Conclusion]The draining for forestation decreased the function of carbon sink from natural marsh wetlands by about 2/3 significantly at first 10 years, then it needs 30 years at least for the carbon sink to be recovered to the same as natural marsh wetlands; So, draining for forestation should be avoided in the temperate marsh wetlands.

Key words: Xiaoxing'an mountains, marsh wetland draining for forestation, Larix gmelinii plantation, greenhouse gas emission, vegetation net carbon sequestration, carbon source or sink of ecosystem