Factor Analysis and Numerical Taxonomy of Soils1
- Rodney J. Arkley2
A large computer and multivariate statistics were used to select a minimal number of soil properties with the greatest prediction value for use in soil classification. Key communality cluster analysis and principal axis factor analysis with Varimax rotation were applied to six sets of soil data containing from 21 to 66 profile characteristics. The analyses selected from four to seven factors or dimensions which were highly independent of each other and contained from two to four highly correlated soil properties. These factors accounted for 100% of the communality and 90% of the raw variance in all properties used in each set. In spite of the diversity of the soils (220 and 620 California, 59 and 86 World, and 148 Ohio soils) the same factors appeared repeatedly; i.e., properties related to reaction, hue and chroma, texture, color value and mottling, and profile differentiation in five analyses and solum thickness in four. Composite factor scores computed for each soil were also used for Numerical Taxonomy. The groupings formed were sufficiently similar to those of the New Classification System of the USDA to suggest that these methods may be very useful in soil classification. The use of these methods for the development of a coordinate system of soil classification is discussed.Please view the pdf by using the Full Text (PDF) link under 'View' to the left.
Copyright © .