Search

Publisher Summary This chapter discusses the objectives and problems of soil classification. The classification of soil is done to understand the relationships among soils. It is assumed that soils fall into natural clusters or groups that can then be ordered into a classification. The chapter discusses the numerical taxonomy or cluster analysis of soils. Numerical taxonomy is the grouping by numerical methods of taxonomic units into taxa based on their character states. An approach is discussed for the examination of taxonomic structure—that is, ordination, which is normally used when the distribution of individuals in n-space tends to be continuous rather than in distinct clusters. Two kinds of ordination have been used in soil studies: Q-type ordination and R-type ordination. The chapter presents three statistical methods for comparing soil classification, including cophenetic correlation, coefficient of association, and Wilk's criterion. Cluster analysis of soils is a most effective means of classifying soils when the number of distinct soils or soil groups is limited as within a relatively small land area.