Evaluation of Computed Tomography to Determine the Distribution of Macropores in Soil

Abstract

Preferential flow paths in some soils result in the rapid movement of water and associated chemicals to groundwater. Macropores are one type of preferential path. Examples of macropores include earthworm tunnels, decayed root channels, shrink/swell cracks, tillage cracks, etc. Water movement through a soil can potentially be modeled if the number, size and extent of the macropores are known. At present no easy and accurate method of characterizing macropores has been developed. X-ray computed tomography (Cf), developed for medical purposes, uses scan techniques to secure multiple views of an object and provides ameans of obtaining nondestructive internal cross sections of objects. The spatial resolution of modern scanners is on the order of I mm, providing a potential means of detecting small air-filled pores within the soil. Undisturbed soil cores, 200 mm in diameter, were taken in cultivated and uncultivated areas. A medical Cf scanner was used to scan these cores at depth intervals varying from 10 mm to 50 mm. Data for each scan were analyzed using microcomputers to display images and to determine the size and number of macropores for each scan image. Cores were physically sectioned at scan locations to visually compare the size, location and continuity of macropores in the sections with those shown in the scan images. Laboratory cores packed with soil were scanned as standards to determine the response and contrast resolution of the Cf scanner for dense material containing sharp density discontinuities. Cores were packed with different types and densities of soil. Artificial "macropores" were formed in some packed cores using glass tubes and holes formed with wire probes to assess the resolution of the scanner. Macropores of various sizes and types were found in all of the field cores scanned with the majority of macropores being associated with earthworm tunnels. The number and size of macropores were approximately the same for all cores below a depth of approximately 50 mm. The number of macropores found above this depth was affected by a large number of roots in the grass-surfaced core and by tillage in the bare-surfaced core. Many macropores were continuous to depths of 600 mm or more. Some passed completely through the core. Dye tests and physical sampling revealed that some of the macropores functioned as preferential flow paths and had very high flow rates. Analysis of the packed cores revealed that the detection of small holes in a dense medium by Cf depends not only on the size of hole but also on the density of the medium. Various attenuation values were examined to determine the optimal threshold value to use for the detection and measurement of macropores in the field cores. The average attenuation value for air in the holes in the cores was not zero as was initially assumed, but varied with the size of hole. The results showed that a CT scanner can accurately determine pore locations and sizes for pores I mm or larger. Accurate determinations of macropore characteristics require that proper values of parameters for the medium and conditions being analyzed be assessed. Use of microcomputers greatly enhanced the ability to display images, analyze features of interest, detect and characterize individual macropores. The main conclusion of the study was that the Cf scanner when coupled with use of computers is a potentially valuable tool for characterizing macropores.