Abstract: Gravel (>2 mm) is one of the main parameters for estimating soil carbon pool. To assess the effects of gravel on soil bulk density (BD) and organic carbon density (SOCD) in Pinus massoniana plantations, we estimated the BD and SOCD at the 0-10, 10-20 and 20-40 cm soil depths of 131 plots under two different conditions, with and without removing gravel. The BD of each soil layer after removing gravel was 0.58-1.57, 0.60-1.67, and 0.59-1.75 g·cm^-3, respectively, which was significantly lower than that before removing gravel. Gravel increased the BD by 6.5%-6.8%. The SOCD of each soil layer before removing gravel was 8.93-65.97, 7.63-59.08, and 8.79-94.53 t·hm^-2, respectively, which was higher than that after removing gravel. Overall, by neglecting the effect of gravel, SOCD was overestimated by 4.9%-11.8%. As gravel content increased, the relative deviation in the estimated BD and SOCD among different methods increased. When the gravel content was higher than 20%, the estimated SOCD at soil layer of 0-40 cm showed a significant difference between neglecting gravel and removing gravel, with the former being 29.7%-47.4% higher than the latter. In conclusion, gravel markedly affected the estimations of BD and SOCD. It was recommended that SOCD should be estimated by the method that not only uses the BD after removing gravel but also considers gravel as a correction factor (especially when gravel content is above 20.0%) to avoid overestimation of soil carbon pool.

Key words: gravel, bulk density, organic carbon density, relative deviation, plantation.