The experiment was conducted at twelve grazing plots denoting four treatments defined along a grazing gradient with three replications: non-grazing (0 sheep/ha, NG), light grazing (0.75 sheep/ha, LG), moderate grazing (1.50 sheep/ha, MG) and heavy grazing (2.25 sheep/ha, HG). Using an automatic cavity ring-down spectrophotometer, we measured CH₄ fluxes from March 1 to April 29 in 2010 and March 2 to April 27 in 2011. According to the status of soil freeze-thaw cycles (positive and negative soil temperatures occurred in alternation), the experiment was divided into periods I and II. Results indicate that mean CH₄ uptake in period I (7.51 µg CH₄–C m⁻² h⁻¹) was significantly lower than uptake in period II (83.07 µg CH₄–C m⁻² h⁻¹). Averaged over 2 years, CH₄ fluxes during the freeze-thaw period were −84.76 µg CH₄–C m⁻² h⁻¹ (NG), −88.76 µg CH₄–C m⁻² h⁻¹ (LG), −64.77 µg CH₄–C m⁻² h⁻¹ (MG) and −28.80 µg CH₄–C m⁻² h⁻¹ (HG).