Estimating soil organic carbon and nitrogen stock based on high-resolution soil databases in a subtropical agricultural area of China

Abstract

Quantifying soil organic carbon stock (SOCS) and total nitrogen stock (STNS) of the agriculture soils in China is crucial for predicting the future of global climate change and planning agricultural management practices. However, most of existing studies were based on relatively coarse soil databases, leading to large discrepancies in the SOC and STN estimates. The present study was conducted in Changting County of Fujian Province in China, covering an area of 29,375ha. A high-resolution soil database at a county-scale of 1:10,000 was employed to analyze the SOC density (SOCD) and STN density (STND) under different soil groups and land use types. Results indicated that the mean SOCD and SOCS for the surface layer (0–15 cm) were 2.93 ± 0.38 kg C m−2 and 861 Gg, respectively, and those for STND and STNS were 0.37 ± 0.06 kg N m−2 and 108 Gg, respectively. The SOCD and STND showed significant disparities among soil groups and land use types. We also compared the SOCD, SOCS, STND and STNS estimated from a coarser 1:1000,000 soil database, which was the most detailed soil database at the national scale, with the estimates in this study. The relative deviations in estimating SOCD and STND by soil groups from the coarser database were 8.1% and 60.6%, respectively, which could be primarily attributed to the assignment errors of soil attributes and polygon area variations in the coarse-scale soil database. Moreover, we found that the dominant factors determining SOCD were topographic and climatic factors (e.g. elevation and mean annual precipitation), while farming practices (e.g. N fertilizer application and irrigation conditions) as well as elevation were more important for STN estimates. Further, the deviations of STN estimates from the coarse- and fine-scale soil database were higher than those of the SOC estimates because: (1) The topographic and climatic factors had slight effects on STND from coarse- and fine-scale soil database while significantly affected SOCD; and (2) STN estimates tended to be more easily influenced by anthropic farming practices (e.g. significant irrigation conditions and N fertilization application) in the coarse-scale database than that in the fine-scale database. Therefore, a high-resolution soil database may improve the accuracy of SOC and STN assessment, especially for STN assessment, and the application of finer soil databases is a critical step for estimating SOCS and STNS at national and even global levels, and thus for accurately assessing regional C budget.