Comparison of soil linear shrinkage curve from extracted cores and in situ

Abstract

In situ measurement of the linear shrinkage curve (LSC) of soil has been restricted so far to highly swelling clay soils because of the limited precision of thickness measurements of the soil layer. This limitation was eliminated by using electronic displacement transducers that allowed a precision of ±10 µ m on the thickness measurements of 20-cm soil layers. The method was further improved by using block kriging to estimate non-destructively and locally the water content of the soil layer at the same spot where the shrinkage was monitored by the shrinkfiswell gauge.

The LSC of 2 weakly swelling soils measured by this method had a shape similar to that of highly swelling soils, although their shrinkage range was 2 orders of magnitude smaller.In situ LSCs of the soil layers were compared with the LSCs of extracted soil cores measured in the laboratory. In situ and laboratory shrinkage curves were significantly different. This discrepancy is thought to be principally due to the rewetting of the soil cores in the laboratory before the shrinkage curve measurement without any mechanical stress applied to their external surfaces.