Spatial Distribution And Storages Estimation Of Soil Organic Carbon And Soil Inorganic Carbon In Xinjiang, China

Arid ecosystems cover47%of total terrestrial area of the Earth, where1/3of global "missing carbon (C) sink" exists. Hence, variation of carbon storage in arid ecosystems could significantly change global climate. Xinjiang is a typical arid and semi-arid region in China, which plays a significant role in the carbon cycle of terrestrial ecosystem in China and accordingly influences the climate. However, the distribution of carbon in soil and the storage in Xinjiang are still unclear to date due to very limited data. This study made a detailed survey of the carbon in Xinjiang and used methods from Pedometrics to investigate the profile distribution of soil organic carbon (SOC) and soil inorganic carbon (SIC).DThen a geospatial model has been developed for predicting the spatial distribution of SOC and SIC by considering factors including terrain features, climate factors, vegetation index and land use. Finally, we estimated the total soil carbon storages in Xinjiang, and investigated the carbon stocks for different land use types at five ecological zones in Xinjiang. Results of this study advance understanding of the carbon cycle in the arid zone ecosystem. The main results of our study were given in the following.(1) The SOC content was generally decreased while SIC content was increased with increasing depth in the soil profile (0-100cm) in Xinjiang. At all ecological zones, the SOC density at layer of10-20cm was lower than those at layers0-10cm and20-40nm. The SOC density reach maximum at layer of20-40nm and was decreased with depth at>40cm. The SOC densities at the Altai, the west Junggar zone (Ⅰ), and the Tianshan Mountain zone (Ⅲ) were higher than those at the other two ecological zones. The differences were particularly significant at the top layer (0-10cm). At the layer of20-100cm, the SOC density of Tianshan mountain zone (Ⅲ) was significantly higher than those of other ecological zones. The SOC at the layer of20-100cm at Tianshan mountain zone (Ⅲ) was the largest contribution for Xinjiang SOC pool. The SIC density was firstly increased in the depth0-40cm and then decreased in the depth40-100cm at all the ecological zones except the Tianshan mountain zone (Ⅲ). The SIC density was highest at the Tarim basin zone (Ⅳ) than those at other ecological zones. The difference was particularly significant below the depth of60cm. The SIC at soil depth below60cm at the Tarim basin zone (Ⅴ) was the largest contribution to the SIC pool in Xinjiang.(2) The SOC and SIC content showed non-normal distribution (P<0.05) in the profile (0-100cm) in Xinjiang. The SOC content differed significantly at different layers, which was firstly decreased and then increased. The SIC content had medium variability so its variation with depth was small. The exponential model can well simulate the spatial distribution of SOC and SIC content in different layers of profile. The coefficient of determination at all layers reached a significant level. With the increase of soil depth, the area with SOC content<5g kg-1was gradually increased, but the area with the SOC content between10-20g kg-1was decreased. The distribution of SIC in Southern Xinjiang was different from that in Northern Xinjiang. Specifically, the SIC content was increased with increasing depth in the Northern Xinjiang, whereas in the Southern Xinjiang the SIC was firstly decreased at the depth0-40cm and then was increased in the depth of40-100cm.(3) Five models [Ordinary Kriging (OK), Multiple Linear Regression (MLR), Regression kriging (RK), Geographically Weighted Regression (GWR), and Geographically Weighted Regression Kriging (GWRK)] have been used to predict the spatial distribution of SOC and SIC density. Results showed that the SOC density in the middle of the Western part of the Tianshan Mountains was≥21kg m-2, which was higher than those at any other area. The SOC density at Eastern and Western of Xinjiang was≤6kg m-2. In fact, most of these areas had SOC density smaller than3kg m-2. These areas had the lowest SOC densities in Xinjiang. The SIC densities in the Eastern and Western of Tarim Basin were greater than24kg m-2, which was higher than those at other regions. In contrast, the SIC density in the Northern (<5kg m-2) was the lowest in Xinjiang. GWR model can perform better than MLR to simulate the variation of SOC/SIC with their explanatory variables. The best models for prediction of SOC and SIC densities are GWR and GWRK, respectively.(4) By investigating the factors influencing spatial distribution of SOC and SIC, we found that the key factors for spatial distribution of SOC was normalized difference vegetation index (NDVI), which controlled all over the Xinjiang region. The key factors influencing the spatial distribution of SIC and their controlled area percentage were annual average precipitation (P)(47.76%), evapotranspiration (ET)(21.38%), annual average temperature (T)(17.36%), slope (S)(9.44%), and land use Composite Index (La)(4.05%).(5) The estimated C stocks was obvious differences based on our quantitative data and qualitative data, which were38.45Pg C and46.60Pg C, respectively. It is estimated that the whole Xinjiang region has46.60Pg C,19.56Pg SOC, and has27.04Pg SIC. The SOC and SIC stocks accounted for41.97%and58.03%of Xinjiang soil C stocks, respectively. The SOC and SIC storages in the five ecological zones are in the order of Ⅲ>Ⅴ> Ⅳ> Ⅰ> Ⅱ and Ⅳ> Ⅲ> Ⅴ> Ⅱ> I, respectively. The soil types of higher SOC stocks were brown caliche soils and brown desert soils, accounted for10.26%and10.05%of the SOC storages, respectively. The aeolian sandy soils and brown desert soils had higher SIC stocks, accounted for29.07%and14.83%of the SIC storages, respectively. The SOC density was highest for gray-cinnamon soils and dark felty soils, while lowest for aeolian sandy soils and skeleton soils. The SIC density was highest for shrubby meadow soils and desert solonchaks, while lowest for grey forest soils and brown coniferous forest soils. The soil C storages for different land use types followed the order of unused land> grassland> forest land> farmland> urban residential land. The storages are proportional to the area of the land types.