Abstract: 【Objectives】Changes in vegetation as a result of converting cultivated land into forested areas are known to effectively prevent the soil erosion as well as significantly increase the soil organic carbon storage in these regions. However, the spatial change of soil CO2 emission and its control mechanism are poorly understood and can thus lead to further uncertainties in the quantitative estimations of soil carbon sequestration in these reforested areas. A typical re-forested hillslope was selected in order to investigate the spatial variation of soil CO2 emissions and its control mechanism in the Loess Plateau. This study aims to provide a scientific basis for further understanding the Loess Plateau organic carbon turnover and improve methods for estimating the carbon balance of terrestrial ecosystems.【Methods】In order to determine tempo-spatial dynamics of soil CO2 emission of sloping cultivated land and its influencing factors, the re-forested hillslope(250 m total length)was divided into 5 sections-hilltop, shoulder, upper, middle and lower slope-and each section analyzed. The point method was used to estimate the vegetation coverage of all study plots selected at intervals of 10 meters along the entire slope. Soil samples were collected by drill and root density, soil organic carbon(SOC)content and 137Cs inventory were analyzed. In situ soil CO2 emission was monitored by LI-8100 carbon flux automatic systems on a monthly basis, and soil water content and soil temperature(at a depth of 5cm)were also measured. Correlation and regression analysis was applied to determine the main factors that affect spatial soil CO2 emissions.【Results】 The results show that the temporal dynamics of soil CO2 emission rates at different slope positions during the data collection period was highest in the summer, followed by autumn, with spring having the lowest observed soil CO2 emission rates. When calculating the average value of soil CO2 emission rates across the whole hillslope, emission rates for summer and autumn were found to be higher by 48% and 9%, respectively, when compared to spring. The spatial patterns of soil CO2 emission rates were found to be similar across spring, summer and autumn and the average emission rate of the three seasons was found to decrease as follows across the slope: hilltop(reference)2.51±0.07 μmol/(m2·s) shoulder2.19±0.17 μmol/(m2·s) lower1.88±0.12 μmol/(m2·s) middle1.71±0.09 μmol/(m2·s) upper1.62±0.12 μmol/(m2·s). Using the hilltop as a reference, the 137Cs inventory in the upper and middle hillslope was lower by 46% and 29%, respectively; however 137Cs inventory calculated at the shoulder and lower region of the hillslope was 88% and 52% higher than the reference. These results indicate that there was serious soil erosion at the upper section of the hillslope with lighter soil erosion at the middle section. Furthermore, soil accumulation occurred at both the shoulder and lower sections, with more significant accumulation occurring at the shoulder. We found that soil CO2 emission rates significantly correlated with the slope gradient(P0.01)and 137Cs inventory(P0.01)during the data collection period. Interestingly, only in summer did the soil CO2 emission rates have significant correlation with soil moisture, soil temperature and SOC stock(P0.01). No significant relationship was found between soil CO2 emission and root density.【Conclusions】These results suggested that soil erosion and deposition processes induced by the change of topographic slope are the main factors controlling the spatial variation of soil CO2 emission rate on the Loess plateau ecological forest slopes. These factors should thus be taken into consideration in the quantitative evaluation of the effectiveness of soil carbon sequestration by the Grain to Green Project.