Characterizing the changes in soil porosity by computed tomography and fractal dimension

Information in regard to nondestructive and repetitive measurements of changes in soil porosity ( epsilon ) that may occur during wetting and drying is limited, particularly information regarding changes on a very small scale. The objectives of this study were (i) to determine whether the changes in epsilon that may occur during wetting and drying at scales as small as 2 x 2 mm can be discriminated by computer-assisted tomography and (ii) to use a published theoretical equation to estimate surface fractal dimension (D) from epsilon and to determine whether D is sensitive to wetting and drying. Computer-assisted tomography was applied to gamma-ray attenuation to measure dry bulk density ( rho ), before and after wetting, at 2 x 2-mm resolution of water-stable soil aggregates (WSA) 2 to 4, 0.71 to 1.40, and 0.25 to 0.71 mm in size and packed separately in acrylic cylinders. Columns with similar particle size were also prepared for unstable soil aggregates (USA). Before wetting, epsilon computed from rho in WSA, ranged from 0.621 to 0.740; after wetting the range was 0.604 to 0.709. In USA, epsilon ranged from 0.489 to 0.562 before wetting and from 0.457 to 0.516 after wetting. The lack of a 1:1 relationship between the before and after wetting data for epsilon indicated there were significant differences between the two. Initial aggregate size (x), wetting, (w), and the interaction w x accounted for 74% of the variability in epsilon of USA compared with only 47% of the variability for WSA. The estimates of D, obtained using the theoretical equation and epsilon , ranged from 2.154 to 2.236 for WSA and from 2.055 to 2.12 for USA. Wetting, x, and w x accounted for 47% of the variability in D of USA compared with 69% for WSA. Pore continuity (PC), estimated using a theoretical relation involving PC, epsilon , and D, decreased from 0.45 to 0.30 after wetting in USA and from 0.60 to 0.55 in WSA.