Quantifying relationships between X-ray imaged macropore structure and hydraulic conductivity

Sammanfattning: Computer based simulation models which estimate hydraulic properties of soil usually make use of statistical approaches such as multiple regressions that relate hydraulic properties to widely measured soil properties like textural fractions, organic matter content and bulk density. The problem is that hydraulic properties are usually only weakly correlated to these properties, because it is more strongly dependent on the characteristics of larger pores in the soil. The characteristics of macropore in soil can be determined by X-ray tomography. The aim of this study was to evaluate the predictive performance of X-ray CT derived macropore characteristics for hydraulic properties of soil. Twenty undisturbed soil cores were sampled (6 cm high, 6.5 cm diameter) from a soil pit, with replicate cores taken at different depths from clay soil in Skuterud catchment, Norway. Hydraulic properties’, including saturated hydraulic conductivity, was measured using constant head method and unsaturated hydraulic conductivity was measured using Mini-Disk Infiltrometer. Soil columns were scanned using X-ray CT scan at a voxel resolution of 40 microns. Minimum automatic threshold algorithm method was chosen to segment the images between pore space and soil matrix. The images were analyzed for quantifying the macropore characteristics using ImageJ, SoilJ, BoneJ and GeoDict. Significant correlations were observed between hydraulic conductivity and most CT derived macropore characteristics. Many of the X-ray CT derived macropore characteristics were interrelated. Best multiple linear equations for predicting the saturated and unsaturated hydraulic conductivity were estimated from macropore characteristics. Among the macropore characteristics critical pore diameter largely contributed to the saturated hydraulic conductivity. Macroporosity and connected macroporosity mainly controlled the unsaturated hydraulic conductivity. Physics-based approaches like percolation models based on critical pore diameter serves better than statistical approaches for decision making in soil and water management.