Abstract: [Objective] Through the comparison among the estimated result of the soil organic carbon density in the spruce (Picea asperata) and red birch (Betula albosinensis) natural forests in the Xinjiashan Forest Farm in Qinling Mountains with different estimation method, we tested the improvement of the new method on estimation accuracy by subtracting the root volume.[Method] The organic carbon densities of the mineral soil layers (top soil layer, core soil layer and subsoil layer) and organic soil layers (non/semi-decomposed layer namely L + F and fully decomposed layer of litter namely H) were estimated separately. For the estimation of the organic carbon density in the mineral soil layer, the estimation accuracy is improved by deducting the root volume of trees based on the existing method. The method for estimating the volume content of roots in each layer contains 4 steps as follow. Firstly, the root biomass of a single tree was estimated with the regression equations proposed by the previous study, which is multiplied by the growth density of the forest to obtain the total biomass of the roots per unit area of the forest land; secondly, by measuring the biomass and the volume of a partial sample of the root system,the density of the partial root sample is calculated to represent the density of the whole root system; then, the total root volume of the forest area per unit area is calculated by dividing the total biomass of the roots per unit area of the root system by the density of the root system; Finally, according to the distribution result of the root system along the depth obtained in previous studies, the total root volume per unit area is distributed to each soil layer and the content of root volume is calculated. For the estimation of the carbon density of the organic soil layer,the volume of the base of the trunk contained in the organic soil layer is subtracted, which is calculated with the mean ground diameter. In addition, due to the extremely uneven distribution of different components in the organic soil layer, the different components in the litter (organic soil layer) are carefully grouped according to the source organs and physical forms, and the organic carbon content of each component is determined.[Result] Thicknesses of top soil layer (A), core soil layer(B) and subsoil layer(C) were 19.10, 14.20, 31.03 cm and 18.57, 15.13, 28.13 cm for the spruce and the red birch, respectively. The SOC contents in A, B and C horizons were (44.56±3.72) g·kg^-1, (25.63±1.77) g·kg^-1 and (10.79±2.28) g·kg^-1 respectively for spruce;they were (34.11±5.46) g·kg^-1, (19.06±4.95) g·kg^-1 and (11.02±3.86) g·kg^-1 respectively for red birch. The carbon contents of various components in organic soil layer of the two stands were significantly different (P < 0.05). The organic carbon contents of spruce cones, moss, tree branches and roots in litters were greater than 600 g·kg^-1. Followed by the organic carbon contents of leaves, which were (458.90±46.81) g·kg^-1 for spruce forest and (420.72±55.66) g·kg^-1 for red birch forest. The organic carbon contents of other undistinguishable fine particles was the minimum which were less than 300 g·kg^-1. The root volumes in A, B and C horizons were 66.81 (3.5%), 20.69 (1.5%) and 9.18 (0.3%) m³·hm^-2 respectively for spruce, and they were 50.57 (2.7%), 31.75 (2.1%) and 17.22 (0.6%) m³·hm^-1 respectively for red birch. The carbon density of organic soil layers was 4.26 kg·m^-2 for spruce, less than that of the original method by 0.73%; H and L+F layers accounted for 84% and 16% respectively; the carbon density of mineral soil layers was 16.58 kg·m^-2, less than that of the original method by 2.13%. The carbon density of organic soil layers was 3.49 kg·m^-2 for red birch, less than that of the original method by 0.48%; H and L + F layers accounted for 90% and 10%; the carbon density of mineral soil layers was 14.06 kg·m^-2 for red birch, less than that of the original method by 1.61%.[Conclusion] The estimated total organic carbon densities of soil in spruce forest and red birch forest decreased by 1.85% and 1.39% respectively after deducting the content of root volume, which suggested that the currently-estimated forest soil carbon storage might be generally higher than the actual value.

Key words: soil organic carbon density, carbon storage estimation, forest soil, natural forest, forest floor, new method