USE OF STEPWISE DISCRIMINANT ANALYSIS TO ASSESS SOIL GENESIS IN A YOUTHFUL SANDY ENVIRONMENT

Publisher Website
Google Scholar
Abstract
To evaluate statistically the age-related changes of youthful sandy soils at a beach ridge and dune complex in northeastern Illinois, I used stepwise discriminant analysis: (1) to determine which pedogenic variables have the greatest ability to discriminate between different aged soils; (2) to evaluate whether pedogenic changes are significantly great enough to classify a majority of pedons into their predetermined groups; and (3) to attempt a quantitative assessment of the relative degree and rate of soil formation. The analyses suggested that a near steady state is attained as the soil approaches an age of about 300 years. The variables that best differentiate the soils were the thickness of the A1 horizon and the amount of clay in the B horizon. Finally, the ability to use discriminant analysis to test the probability of each sampling location belonging to a particular age is warranted because 41 of 60 sampling locations were correctly classified into their a priori age group. To evaluate statistically the age-related changes of youthful sandy soils at a beach ridge and dune complex in northeastern Illinois, I used stepwise discriminant analysis: (1) to determine which pedogenic variables have the greatest ability to discriminate between different aged soils; (2) to evaluate whether pedogenic changes are significantly great enough to classify a majority of pedons into their predetermined groups; and (3) to attempt a quantitative assessment of the relative degree and rate of soil formation. The analyses suggested that a near steady state is attained as the soil approaches an age of about 300 years. The variables that best differentiate the soils were the thickness of the A1 horizon and the amount of clay in the B horizon. Finally, the ability to use discriminant analysis to test the probability of each sampling location belonging to a particular age is warranted because 41 of 60 sampling locations were correctly classified into their a priori age group. © Williams & Wilkins 1980. All Rights Reserved.

This publication has 4 references indexed in Scilit: