A strand plain soil development sequence in Northern Michigan,USA

A practical problem limiting the theoretical application of a soil chronosequence study lies in the difficulty of identifying an appropriate geomorphic age sequence with more than a few surfaces. A sequence of approximately 75 strand plain beach ridges located along the northern shore of Lake Michigan, USA, overcomes this problem, and was used to explore rates of soil development in the initial stages of podzolization. Well drained pedons were sampled on 24 of the ridges, with an age range of 10 to 5200 years. Chronofunctions were developed using statistical regression techniques. For most forms of extractable Fe and Al, both linear and log–linear functions adequately describe the changes in property with surface age. Dithionite–citrate extractable Fe content, however, shows little correlation with surface age. High r² values (>0.7) for chronofunctions representing most extractable sesquioxide forms suggest that B horizon sesquioxide content increases with surface age, as expected for soils undergoing podzolization. Visual inspection of chronofunction data suggests that scatter increases on surfaces older than 3000 years, possibly because soil spatial variability increases with surface age.Theory suggests that the log–linear form of chronofunctions may be preferable to the linear form for these pedons. In general, equation slopes are steeper for chronofunctions representing only the initial 1650 years of soil development than for those representing the entire 5200-year sequence, suggesting that initially rapid pedogenesis slows over time. Most chronofunctions for pH and weighted profile organic C are not statistically significant, possibly because the temporal resolution of the chronosequence is too coarse to detect valid trends for properties that rapidly approach a steady state.