To investigate the short-term effects of grassland renovation on carbon dioxide CO 2 exchange of intensively managed grasslands in Japan, we conducted CO 2 flux measurements during the renovation process by using the eddy covariance and closed chamber methods in 2007, 2012, and 2013. The flux measurements were conducted at three grassland fi elds: two fi elds used for cutting, one receiving only chemical fertilizer CF and the other receiving composted cattle manure CM annually, and one fi eld used for grazing GM . Chamber measurements revealed fl ushes of CO 2 after plowing 14.45 -23.94 μmol m 2 s 1 and subsequent disk harrowing 7.36 μmol m 2 s 1 , in CF followed by rapid drops of CO 2 flux, which were local and temporary phenomena. The mean CO 2 losses during renovation periods were calculated as 4.52 -4.74, 6.00 -6.81, and 4.57 g C m 2 d 1 in CF, CM, and GM, respectively; CF and CM calculation is based on eddy covariance measurements, including temporal fl ux variations and representing footprint areas. The amount of carbon input during renovation including non-harvested grass biomass stubble and roots and applied manure was estimated as 2.24 -3.50 and 6.17 -8.77 Mg C ha 1 in CF and CM, respectively. Among them, carbon derived from plowed roots and manure is presumably resistant in soil, contributing to long-term soil organic carbon SOC accumulation. Our results also indicate that grassland renovation work does not affect short-term net CO 2 loss signifi cantly, although it affects CO 2 emissions to a certain extent. We can therefore say that net CO 2 loss during renovation is mainly brought by the absence of vegetation in this site. Grasses contribute SOC accumulation through biomass allocation belowground; and it is thus recommended to shorten grassland renovation period, the duration without photosynthesis, to reduce CO 2 loss associated with grassland renovation.