Fragipan formation in loess soils: development of the Bryant hydroconsolidation hypothesis

Abstract

The formation of fragipans is still a matter for investigation and discussion. Bryant (Soil Sci. Soc. Am. Spec. Publ. 24, 1989, pp. 141–150) suggested that the `hydroconsolidation' process (structure collapse when loaded and wetted) may contribute to fragipan formation. Experimental studies of the collapse behaviour of both undisturbed and model loess soils are reported. These demonstrate that the `small clay' model of hydroconsolidation advanced by Rogers et al. [Eng. Geol., 37 (1994) 83–113] appears to explain many aspects of the collapse of loess soils when loaded and wetted. This `small clay' model can reconcile the Bryant suggestion with collapse observations, and in so doing can satisfy all of the important criteria governing the formation of fragipans. In particular it can explain the constant depth of occurrence. The `small clay' model requires the soil structure to collapse under the control of a critical and limited clay mineral content. Too much or too little clay does not allow collapse to occur, as indicated by oedometer test data for both undisturbed and reconstituted loess soils under simulated field conditions. The fragipan should form in a metastable soil, one capable of some linear collapse.