Variability in soil properties and its consequences for design

Abstract

It has been observed that when samples from one particular layer of one particular test pit are analysed by different laboratories the results may vary over an alarmingly large range. Jakobsz and Day (2007) suggested that materials testing laboratories may be slipshod in their testing procedures. There is no doubt that in some cases testing is not done as rigorously as it should be, but this does not appear to explain the size of discrepancies in all cases. Badenhorst et al (2015) suggested that differences in methods of prepa-ration between laboratories may be a significant cause. Preparation can certainly lead to significant differ-ences (Blight 2012, Stott and Theron 2015a), but will also not convincingly explain some of the discrepan-cies. Many soil properties are intimately linked to suction (water retention) potential. A procedure for assessing suction potential of soil samples with very little preparation and minimal disturbance to micro-structure was described by Stott and Thereon (2015b). This technique allows reasonably rapid testing of a number of sam-ples simultaneously with almost no variation in preparation or testing procedure. Tests performed on a range of soils suggest that some show very small variability, whereas others show extremely large difference in suc-tion potential over a very small spatial range. It seems possible that inconsistent results from testing laborato-ries could be an indication that some soils cannot be reliably assessed with some of the popular procedures of current practice. A reappraisal of design philosophy may need to be considered in some cases, particularly in the case of expansive clays.