Characterization of soil aggregate microstructure under different revegetation types using micro-computed tomography

Abstract: Soil aggregate microstructure is a crucial factor that affects various soil physiochemical and biological processes. Vegetation restoration is expected to improve soil microstructure, yet little is known about the extent of changes in soil aggregate microstructure. X-ray micro-computed tomography (micro-CT), in combination with image analysis techniques, can provide three-dimensional (3D) data of porosity and pore size distribution, and is therefore useful to better investigate the 3D microstructure of soil aggregates. The objective of this study was to evaluate the impacts of different revegetation types (artificial shrubland, natural grassland and slope cropland) on the aggregate microstructure on the Loess Plateau using X-ray micro-CT measurements. The sampling sites were all located close to the top of the loess mounds with little difference in terms of aspect, slope gradient, elevation, and previous farming practices. An area of 100 m ×100 m was randomly selected for each site and within this area 3 20 m × 20 m plots were selected for sampling. Three samples of soil aggregate (3-5 mm) within 0.20 m soil profiles selected from each site were scanned at 3.25 μm voxel resolution with SSRF (Shanghai Synchrotron Radiation Facility). Segmentation of the grayscale slices was performed using automatic Otsu thresholding algorithm, and the threshold values were carefully chosen based on visual observation. To avoid edge effects, the sub-volume of 500 × 500 × 500 voxel (1.625 mm × 1.625 mm × 1.625 mm) was extracted for further analyses, representing an inscribed cube of the aggregate. The 3D pore structure was constructed and quantified using the digital image analysis software ImageJ. Aggregate water stability and other soil properties were also evaluated. Results showed that soil organic carbon and aggregate water stability were significantly increased, while soil bulk density was significantly decreased under both revegetation types (artificial shrubland and natural grassland). The surface soil texture showed no significant difference for different revegetation types. The total porosity of the aggregates was increased by approximately 60% and 20%, but the pore number was decreased by about 62% and 68% respectively in the artificial shrubland and the natural grassland compared to the slope cropland. Vegetation restoration affected the pore distribution of soil aggregates, which on the one hand promoted the proportion of >100 μm large pores and on the other hand decreased that of the pore size classes of <30 and 30-75 μm. Revegetation significantly changed the pore shape of soil aggregates, with a shift from regular and irregular pores to elongated pores. The fraction of elongated pores was dominant in all soil samples (on average 81%), and the order was artificial shrubland > natural grassland > cropland. The 3D fractal dimension and connectivity of soil aggregates showed a higher value in both revegetation types, suggesting pore system was improved after vegetation restoration. The total porosity, macro-porosity (>100 μm), fraction of elongated pores, fractal dimension and pore connectivity were significantly higher in the artificial shrubland aggregates compared to the natural grassland aggregates, indicating that the soil structure of the artificial shrubland was more developed than that of the natural grassland. That may be due to a higher organic carbon content and more developed root system under the artificial shrubland. The 3D fractal dimension of soil aggregates showed a curvilinear positive correlation with total porosity, however, Euler number showed a curvilinear negative correlation with total porosity. The fractal dimension and connectivity showed a high sensitivity to the change of soil structure, and thus could be used for evaluating the soil quality during the revegetation in this region. These results from this study can help understand the soil processes and may be used to quantify the effects of management on environment. With the development of computed tomography, it should be widely used to investigate the soil microstructure in more regions; and moreover the process mechanisms of soil aggregates also require further investigation.