Multi-Scale Characterization of Spatial Variability of Soil Organic Carbon in a Semiarid Zone in Northern China

Abstract

The variation of soil organic carbon (SOC) spatial distribution is dependent on the relative contributions of different environmental factors, and the dominant factors change according to study scales. Here, geostatistical and remote sensing techniques were used to gain deep knowledge about SOC spatial distribution patterns and their dominant determinants at different study scales; specifically, the structure of the spatial variability of SOC content at the county, regional, and watershed scales in Aohan, China were analyzed. The results show that altitude and normalized difference vegetation index (NDVI) are the key predictors explaining 49.6% of the SOC variability at the county scale; NDVI and slope are the key predictors explaining 36.2% of the SOC variability at the regional scale; and terrain factors are the most significant factors at the watershed scale. These three scales have a moderate spatial correlation in terms of SOC content. As the study scale widens, the spatial variability attributable to the random factors increases gradually, whereas the variability attributable to the structural factors gradually weakens. Soil type and land use type are the key factors influencing the SOC content at these three scales. At all scales, the SOC contents of the different land use types differ significantly in the order forestland > shrubland > grassland. Conservation of regional soil and water and prevention of soil desertification are effective measures for improving SOC content.