Soil respiration is an important ecological process in which CO₂ from different depths of the soil is transported to the soil surface and released to the atmosphere. The CO₂ production and diffusion rates at different depths in the soil significantly affect the CO₂ flux from the soil surface. Rainfall is a frequent weather event that causes an increase in the soil water content (SWC), and temperature changes induce dramatic variations in soil CO₂ production and transport. To clarify diurnal variations in soil CO₂ concentrations and the response to short-term rainfall, we selected Sejila Mountain in Southeastern Tibet as a model area. Under natural rainfall conditions, the characteristics of abies georgei var smithii forest soil CO₂ concentrations at different depths were analyzed. The results showed that the soil CO₂ concentration increased with soil depth in the following order 60 cm > 40 cm > 20 cm > 10 cm > 5 cm. Mean values were 5365, 5304, 4990, 4335 and 4179 μmol/mol, respectively. There was a significant logarithmic relationship between the CO₂ concentration and the soil depth, given by y=531.1lnD + 3274 (R=0.9764,P=0.004). The soil CO₂ concentrations at 5 cm and 10 cm showed obvious diurnal variations, and could be expressed as a single-peak curve, while diurnal variations at 20 cm, 40 cm and 60 cm were very small. Short-term rainfall significantly affected the CO₂ concentrations in the 5 cm and 10 cm soil layers; the CO₂ concentration in the 5 cm soil layer decreased from 4196 μmol/mol to 4176 μmol/mol at the rainfall stage, and the CO₂ concentration in the 10 cm soil layer increased significantly from 4252 μmol/mol to 4354 μmol/mol. But there was a significant increase in CO₂ in the 5 cm soil layer and a decreased in CO₂ in the 10 cm soil layer during soil water redistribution. Closure of the CO₂ transport channel by water may be an important mechanism governing the increase in the CO₂ concentration in soil layers below 5 cm. At the rainfall and water redistribution stage, there was a significantly negative correlation between the CO₂ concentrations at 5 cm and 10 cm (P < 0.001), but a significantly positive correlation at 10 cm, 20 cm, 40 cm and 60 cm (P < 0.01). The SWC of the 5 cm layer had a significant effect on the CO₂ concentration in the 0-60 cm soil profile. There was a significant linear positive correlation (P < 0.001) between the SWC at 5 cm and the CO₂ concentration profile at the rainfall stage, while at the water redistribution stage it followed a power function model, and had a significantly negative correlation (P < 0.001). In other words, rainfall caused a rapid increase in the SWC of the 5 cm layer, which induced a significant increase in CO₂ in the 0-60 cm profile. The decrease in the SWC of the 5 cm layer at the water redistribution stage promoted a release of CO₂ from the soil into the atmosphere from the soil profile and caused the CO₂ content to decrease. The effects of soil temperature and SWC on the CO₂ concentration were not consistent among the different layers in the 0-60 cm soil profile, which showed that the other layers had the opposite effect, except at a depth of 40 cm. Overall, the results of this study indicate that the CO₂ emission flux induced in the soil by short-term rainfall decreased; however, the CO₂ concentration increased in the layer below 5 cm. This study suggests that diurnal variations in the CO₂ concentrations in the soil are affected by short-term rainfall.