Abstract: Bamboo forest, as a special forest type in China, has a great carbon sequestration potential. However, there are large uncertainties in the estimation of its carbon pools, which is likely related to the heterogeneous distribution of bamboo forest. Few studies were conducted on the spatial heterogeneity of soil carbon, compared with the estimation of vegetation carbon pool of bamboo forests. In this study, we examined the spatial variation of soil organic carbon (SOC) of bamboo forests in Changning County, Sichuan Province using geostatistics (Kriging interpolation method). The spatial autocorrelation of SOC density was low in 0-20 cm soil layer, and medium both in 20-40 cm and 40-60 cm soil layers, showing that spatial autocorrelation gradually increased with increasing soil depth. The optimal Kriging interpolation neighborhood of SOC density was 1.5 km. The SOC pool (0-60 cm depth) of bamboo forests in this county was 2.45 Tg, showing a decreasing trend with increasing soil depth. Spatially, SOC density with high values showed a patchy distribution in southern Changning, and a mosaic distribution in the northern part, thus showing a decreasing trend from south to north. Correlation analysis indicated that spatial variation of SOC was more likely attributed to total nitrogen, soil moisture and vegetation index (a reflection of vegetation growth). Our results can provide an important basis for improving the accurate estimation of SOC, as well as sampling design and forest management for bamboo forests.

Key words: soil carbon storage, spatial variability, geostatistics, interpolation neighborhood, bamboo forest.