Bulk samples from the project soils were fractionated into the readily- or water-dispersible and chemically-dispersible clay (Na2CO3) fractions and concentrated for chemical and mineralogical characterization, the latter of which was further fractionated into coarse (2-0.2µm) and fine (<0.2µm) clay using centrifugation. In comparing the initial screening data to the results from the final project soils, air-drying generally reduced the dispersibility of a given soil pedon. The project soils were also characterized in terms of a variety of physical and chemical properties, such as particle size distribution and soil pH, using standard methods. Surface horizons were generally more dispersive than B horizons, suggesting that organic matter increased the net-charge and dispersibility of surface horizon materials while Fe and Al oxides reduced the dispersibility of the subsurface horizons. In comparison with distilled water, dilute Na solutions generally increased the dispersibility of soil materials presumably because of an increase in the sodium adsorption ratio (SAR) and the presence of loosely held sodium on cation exchange sites. Dispersibility of subsurface horizons, particularly Ultisols and Oxisols, appears to be governed by the content and surface properties of the Fe and Al oxides, as indicated by subsequent statistical analysis, and is therefore sensitive to specific adsorption reactions of calcium and sulfate. Colloid-associated organic matter contents were generally much higher for the WDC compared to the bulk soil and generally confounded quantitative thermal mineralogical analysis.
Correlation analysis and an examination of the correlation matrices for the entire project data set and select groupings of the project data (i.e. surface vs. subsoil horizons, soil orders, mineralogy types, etc.) were used to: (1) identify parameters indicative of soil dispersion behavior, (2) detect potential data outliers, and (3) identify obvious data trends. For the soils which produced significant WDC, better statistical correlations were generally observed for the elemental composition of the easily dispersed fraction and the coarse clay fraction (2-0.2 µm), rather than the fine clay (<0.2µm). The same relationship was qualitatively observed for the x-ray diffraction results.
A step-wise regression analysis was performed on the WDC results using a set of 22 of the 150 measured parameters. When all soil samples were evaluated, including those that produced little WDC, generally poor correlations were observed presumably due to the wide variations of soil types included in the project. Therefore, regression analysis was also conducted on select groupings of the project soils: (1) Ultisols and Oxisols, all horizons; (2) Ultisol and Oxisol surface horizons; (3) Alfisols, Entisols, Inceptisols and Mollisols, all horizons; and (4) Alfisols, Entisols, Inceptisols and Mollisols surface horizons. For the Ultisols, CDB extractable Al and Fe negatively correlated with WDC, especially the surface horizons. Generally poor correlations were observed with factors such as SAR and suspension Na concentration. For the less weathered soil groups, SAR, soil pH and CDB-extractable Fe and Al were identified as factors that could explain a great deal of the variation between project soils.
Last Modified: September 30, 1998
Document Prepared by:
North Carolina Agricultural Research Service
North Carolina State University