NUMERICAL ANALYSIS OF A CHRONOSEQUENCE INCLUDING AN ASSESSMENT OF VARIABILITY

Abstract

One hundred and eighty-three soil samples from 60 pits were collected from a sequence of moraines located in front of the Robson Glacier, Mt. Robson, British Columbia. Univariate analysis of variance, oblique rotation factor analysis, and nonlinear regression were employed to ascertain the major processes operating on the moraines and to determine whether expected trends were masked by the variability of the parent material. Soil properties which varied significantly with depth and moraine age included organic carbon, nitrogen, and pH. Assumed to be functionally related, these properties were also identified with one of two significant factors from the factor analysis. The values for organic carbon and nitrogen were fitted to nonlinear equations involving an exponential term for depth and a logistic term for time; the resulting correlation coefficients were 0.89 and 0.88, respectively. Percent calcium carbonate equivalent, percent iron plus aluminum by citrate-bicarbonate-dithionite extraction, percent iron plus aluminum by pyrophosphate extraction, percent sand, and percent clay showed generally little or no relationship to depth and age. These properties were related to the second factor from the factor analysis, which was identified with water infiltration and abiotically induced carbonation. The absence of stronger depth and age trends for these properties was considered a result of a high degree of variability within the parent material. Key words: Chronosequence, chronofunction, pedogenesis, soil development, soil sequence, soil nitrogen