Spatial Variation And Simulation Of Soil Organic Carbon In Key Belt On The Loess Plateau Of China

The Loess Plateau of China(LP)is the largest and most concentrated loess area in the world,and its deep loess deposit contains abound geoscience information.Soil organic carbon(SOC)is an important part of terrestrial ecosystems and one of the key elements controlling soil fertility and ecosystem function,and plays an important role in global climate change.In order to clarify the spatial variability and influencing factors of SOC in LP critical zone,a ca.860 km south-north transect with a clear environmental gradient was established in the LP.Both classical statistical and geostatistical methods were used to analyze the spatial distribution and variation of SOC in the 0~5 m profile and its relationship with environmental factors.The spatial distribution characteristics of SOC in different soil layers was simulated by using state-space approach,linear regression and pedotransfer function.Soil core samples from the surface to the bedrock were collected in the small watersheds of Changwu Wangdonggou and Ansai Zhifanggou and the vertical differentiation characteristics of SOC under different loess topography in the LP critical zone were analyzed.The results showed:(1)SOC gradually reduced from the south to the north of the transect and SOC gradually decreased with the increase of soil depth in the 0~100 cm profile.SOC in different layers of 0~500 cm profile presented moderate variation with a range of 60~270 km.The spatial distribution of SOC at 0~40 cm layer is mainly affected by precipitation and slope,and the spatial distribution of SOC in the 40~500 cm layer is mainly affected by soil texture and climate.This result indicating that the regional spatial distribution of SOC in shallow soil layer depends on climate and topographic conditions,while the SOC in the deep layer is mainly affected by soil texture and climate.(2)The SOC reserves of 0~500 cm profile in the transect of the LP have obvious spatial distribution patterns,and generally decrease from south to north of the transect.The SOC reserves in different soil layers at regional scale were moderately varied.The main influencing factors of regional SOC reserves distribution are precipitation,temperature,seasonal distribution characteristics of precipitation and slope.The multiple linear regression equation established based on the above 4 variables can explain 46-73% of the spatial variation of SOC reserves in 0~300cm layer.(3)The results of Person correlation analysis and principal component analysis showed that clay content,slope gradient,mean annual precipitation,mean annual temperature and seasonal distribution characteristics of precipitation were the main factors affecting spatial variation of organic carbon concentration in the loess plateau.The results of autocorrelation and interactive correlation analysis show that these 5 variables have good spatial autocorrelation and interactive correlation within a lag distance of 17 km,which can be used for state-space simulation.Different methods were used to simulate the spatial variation of SOC in the transect,and the results showed that the state-space approach had the best performance compared with the multiple linear regression and the transfer function,which could be a good choice for predicting the spatial distribution of SOC in the LP.(4)Generally,the contents and reserves of SOC gradually decrease with the increase of soil depth in Changwu and Ansai.The distribution dispersion of SOC in the upper layer is large,and the distribution characteristics of SOC content and reserves are consistent.The content and reserves of SOC in different soil layers in Changwu are greater than in Ansai.However,the variation degree in Changwu are smaller than that in Ansai.The contents and reserves of SOC in Changwu and Ansai varied moderately.Soil water content is the main factor affecting the difference of SOC content between the two watersheds.