FRACTAL DIMENSION AND LACUNARITY OF BULK-DENSITY DETERMINED WITH X-RAY COMPUTED-TOMOGRAPHY

Quantification of soil structure is important for understanding soil b ehavior. The objectives of this study were to: (i) introduce the use o f the point-distribution method (PDM) for calculating fractal paramete rs of soil density (rho(b)), and (ii) introduce the use of fractal lac unarity in conjunction with the fractal dimension for quantifying smal l-scale soil structure data collected with x-ray computed tomography. Undisturbed silt loam soil cores were taken from forest and grassland sites. Additional silt loam soil collected from a nearby cultivated he ld was sieved and repacked in 76-mm-diam. cores. Four cores were packe d with soil passing through a 2-mm sieve but retained on a 1-mm sieve (1-2 mm). Four additional cores were packed with the soil passing thro ugh a 1-mm sieve (<1 mm). Computed tomography (CT) was used to measure small-scale rho(b) of each 0.1 by 0.1 by 2 mm volume element (voxel) of a slice through a soil core. Calculation of fractal dimension and l acunarity was accomplished using the PDM on CT rho(b) data. Results sh ow that CT rho(b) data are fractal when analyzed using the PDM. Mean f ractal dimension (+/-SD) of the forest, grassland, <1 mm, and 1- to 2- mm groups were 2.42 +/- 0.13, 2.35 +/- 0.14, 2.50 +/- 0.025, and 2.47 +/- 0.013, respectively. Corresponding mean fractal lacunarity of thes e four groups at a voxel size of 3.1 mm were 0.24 +/- 0.025, 0.24 +/- 0.019, 0.16 +/- 0.005, and 0.20 +/- 0.012. Fractal lacunarity, which r eflects the second-order statistics of rho(b), or the uniformity of po res and aggregates within a soil core, is essential for discriminating soils of similar but slightly different structure.