Precipitation drives the accumulation of soil organic carbon in the sandy desert of the Junggar Basin, Northwest China

Accurate quantitative estimation of soil organic carbon (SOC) and its storage is crucial for ecosystem response over desert areas, and is an important component of understanding dryland climate change, which can effectively reflect the processes and regulatory mechanisms in the carbon cycle. Nevertheless, the relationship between carbon sources and carbon sinks in sandy soils is a hotly debated subject. In particular, how plants and environmental conditions interact to affect SOC remains unclear in sandy deserts. Here, we investigated 77 sites at depths of 0-100 cm (soil depth was divided by 0-5 cm, 5-10 cm, 10-20 cm,20-30 cm, 30-50 cm, 50-70 cm and 70-100 cm intervals), collected 1617 soil samples, and obtained the climatic conditions, vegetation types, and edaphic factors associated with the sampling sites. This study aims to determine whether there are divisions in the allocation strategies of SOC by various vegetation types, and to analyze the mechanisms of environmental factors on the soil carbon pool in sandy deserts. Our results found that (1) the mean value of SOC in the sandy desert was 2.19 g.kg(-1), which differed significantly among different vegetation types (P < 0.05), yet there were no significant changes in aboveground biomass and litter biomass. (2) SOC did not significantly differ among soil layers except for the Ceratoides latent (P = 0.04), and Seriphidium terrae-albae (P = 0.02) communities. (3) The mean value of soil organic carbon density (SOCD) was 3.38 kg.m(-2) and increased exponentially with soil depth (R-2 = 0.23, P < 0.05). (4) Mean annual precipitation controlled the spatial distribution of SOC variation in sandy desert. Vegetation types, soil variables, topographic factors did not control SOC, but had a strong interactive effect with climatic factors. The explanatory power of interaction effects often showed an improvement over single variables, manifested as nonlinear enhancement and bivariate enhancement. Our results highlighted the nature of the desert's position in the global carbon cycle and provide key information on missing carbon sinks in the driest zone.