Abstract:Aims Because of its highly diverse flora and unique climate, Xishuangbanna tropical seasonal rain forest (XTSRF) is one of few areas in China characterized by well-protected primary forest types considered to be tropical rainforest. Compared with other rainforests around the world, XTSRF occurs at higher latitude and altitude. Our aim was to investigate variation of CO2 concentration at different time scales under this unique climate and special geographical situation and determine how environmental variables formed the temporal pattern of CO2 concentration. Methods CO2 concentration in XTSRF was measured continuously by infrared gas analyzers (Li-7500, Li-Cor, Lincoln, NE, USA) and recorded every 30 min with a data-logger (CR5000); meteorological variables were also measured and recorded. The static chamber method was used to measure soil respiration weekly. Leaf area index was measured in the middle of every month using a forest canopy analyzer (LAI-2000). Important findings The diurnal pattern of CO2 concentration above canopy (AC) was consistent through the year, with higher values at daytime and lower values at night due to photosynthesis uptake and respiration release, respectively. In contrast, near the forest floor (NF) diurnal patterns were bimodal in most months, with a second peak in late afternoon. Average variation of CO2 concentration in a year was larger at AC (622.8-686.5 mg·m-3) than at NF (659-700 mg·m-3). Difference of CO2 concentration between AC and NF was higher in the rainy season than in the dry season, with the largest value (-62.9 mg·m -3) in October and the smallest (-8.4 mg·m-3) in December. The “flushing" phenomenon in the early morning was primarily affected by stability of the atmospheric boundary layer. In long foggy days and under low wind velocities, forests in the valley had little air circulation, but local circulation resulted in sharp increases of CO2 concentration in the afternoon. CO2 concentration of NF and AC were dominated by soil respiration and canopy metabolism, respectively.