Saturday, 15 July 2006

Quantification of Soil Micromorphology from Images of Soil Sections.

Daniel Gimenez1, Adolfo Posadas2, and Hyen Chung Chun1. (1) Rutgers Univ, Dept of Environmental Sciences, 14 College Farm Rd., New Brunswick, NJ 08901-8551, (2) International Potato Center, Av. La Molina 1895; PO Box 1558, Lima - Peru, Peru

The use of soil sections has a long tradition in soil science. Important libraries of thin sections around the world contain relevant information to interpret soil properties and processes. Faster and affordable computers together with theoretical developments make that information more accessible than ever. Among the techniques available for quantifying soil structure from 2-D patterns, the Multifractal Spectrum (MS) and Normalized Entropy (NE) are particularly promising. Despite the potential of these techniques for quantifying soil micromorphology, there have not been extensively used in soil science probably because the interpretation of the results is complicated by the presence of multiple parameters. Adding to the uncertainty is the fact that the techniques can be used to characterize either pore size distribution (without considering their location) or the spatial distribution of pores (without explicit consideration of their size) from soil images. The difference is not trivial considering that MS and NE parameters can be used to predict soil processes. There is a need for a systematic investigation on links among MS and NE results and basic soil properties, in particular with field description of soil morphology. The objectives of this work were to use a large data set of micromorphological images to: 1) compare results of analyses when based on pore size distribution and on pore spatial location, 2) correlate parameters from multifractal and normalized entropy techniques with basic soil properties and field description of soil morphology, and 3) formulate relatively simple but sensitive indices of soil heterogeneity based on parameters from these functions. Soil sections from different sources were digitized and the MS and NE calculated from each image. In a selected group of soil images the same functions were calculated at several resolutions. Both techniques separated images in distinctively different groups, in same case clearly related to soil structure type. Similarly, differences in the structure of soil aggregates of three sizes were manifested across soils and management types. Based on the results of this work we suggest that pore microstructure can be quantified and results interpreted in the context of soil properties and processes.

Back to 2.1B Soil Hydrology, Structure, and Micromorphic Propoerties (Soil Porous System) - Poster
Back to WCSS

Back to The 18th World Congress of Soil Science (July 9-15, 2006)