Soil CO2 emissions from different slope gradients and positions in the semiarid Loess Plateau of China

Abstract Knowledge of CO 2 emissions under different slope gradients and positions and its controlling factors is critical in accurately estimating CO 2 emissions and carbon cycling on the slopes of eroded regions. In this study, three east-facing plots of 100 m 2 (20 m × 5 m) with a slope gradient of 0.5° (S 0.5 ), 1° (S 1 ), and 3° (S 3 ) were established in an eroded gully of the semi-arid Loess Plateau, China. The CO 2 emission, temperature, moisture, runoff, sediment, fine root biomass and grain yield of these three plots were measured from October 2013 to September 2015 to investigate the relationship between slope gradients and soil CO 2 emissions. The results showed that the mean annual cumulative CO 2 emissions at S 1 and S 3 (731.0 ± 65.1 and 628.3 ± 74.8 g C m −2  year −1 ) were about 13.4% and 25.5% lower than that at S 0.5 (843.7 ± 84.9 g C m −2  year −1 ). The CO 2 emissions were higher at bottom slope than at upper slope, with an increase of 26.2% at S 3 , 22.9% at S 1 and 14.5% at S 0.5 , respectively. The mean soil moisture ranged from 40.8% to 44.8% water-filled pore space (WFPS) among the slope gradients, and from 35.8% to 45.6% WFPS among the slope positions. There was a significant difference in mean fine root biomass among different slope gradients (S 0.5  > S 1  > S 3, P   0.05), but no significant difference among different slope positions. The mean soil organic carbon (SOC) ranged from 8.8 g kg −1 at S 3 to 9.9 g kg −1 at S 0.5, and that at the bottom and middle slope were higher than that at the upper slope at S 1 and S 3. Slope differentiated soil moisture content and redistribution, and the thus derived spatial differences in fine root biomass and crop yields, was the major factor influencing the soil CO 2 emissions among slope gradients and positions. Slope gradients and positions should be considered when estimating soil CO 2 emissions and carbon cycling in the complex and fragmented topography regions.