Soil-age relationships and correlations: comparison of chronosequences in the Ljubljana Basin, Slovenia and USA

Abstract

A comparison of soil development indices established in chronosequences from the Ljubljana Basin (humid climate), California (Mediterranean), Pennsylvania (humid) and New Mexico (arid) suggests that the rates of soil development expressed by the indices increase with increasing precipitation, even though some individual soil properties, notably rubification, develop faster in xeric areas. Parent material composition does not seem to affect the indices significantly. Therefore similarity of present climate and climatic history between compared chronosequences remains a pre-requisite for soil-age correlations. The chronofunctions are not suitable for exact dating and age correlations because variability of soil development within each geomorphic surface prevents the dependent variables (soil properties) being used to predict the independent variable (soil age). Changes in age estimates affect calculated rates of soil development. In the Ljubljana Basin chronosequence rates changed by up to 20% for semi-logarithmic chronofunctions and by up to 180% for linear chronofunctions if minimum or maximum ages were used instead of best estimates. Better dating methods and more replicate profiles on each geomorphic surface would diminish these problems. Statistical techniques that account for variability in both variables (maximum likelihood estimates) also help improve age correlations. Within the Ljubljana Basin chronosequence some time dependent soil properties can be used to discriminate between groups of geomorphic surfaces, but each of them has a limited ability to do so because the confidence intervals for deposits overlap. The inability of soil properties to discriminate between all members of the chronosequence is attributed to the large intrinsic variability of soils, the insufficient number of replicate profiles on each deposit and the decreasing rate of soil development with age.