Spatial and Temporal Distribution of Desiccation Cracks in Shrink-Swell Soils

Abstract

Soil crack volume estimates, which are important for hydrology models on shrink-swell soils, are currently based on field measurements of vertical shrinkage and an assumption of isotropic shrinkage; however, few studies have validated the resulting crack volume estimates and studies have been limited to soils with very high shrink-swell potentials. In addition, the spatial variability of soil cracking potential is not well understood. First, I was able to improve in situ measurements of soil shrinkage by using a single borehole for all vertical soil movement and water content measurements. Then measurements of soil layer thickness and water content were made for seven soils with varying COLE values, from 0.01 to 0.17 m m^(-1). Soil crack volume was estimated using cement slurry and photographing excavated soil layers at the end of the study. Over drying and wetting cycles, the relationship between soil layer thickness and water content was linear. Modifying an existing crack volume equation with shrink-swell potential and water content was a better fit to cement-estimated crack volume than the unmodified estimates, improving the r^(2) from 0.06 to 0.84. The model over-predicted soil crack volume by a factor of 10 and a minimum shrinkage volume was required to generate visible soil crack volume. Finally, proximally-sensed bulk apparent electrical conductivity was highly correlated to inorganic C, and the depth of maximum sensitivity of the instrument was deeper than suggested by previous research in coarser textured soils. Because inorganic C is related to shrink-swell potential, it may be possible to use proximal sensors to map shrink-swell potential variability.