Predicting land surface movement based on experimentally determined soil shrinkage behaviour

Abstract

Understanding and predicting shallow subsurface deformation and resulting vertical land surface movement due to shrink-swell behaviour of clayey soils is crucial for infrastructure resilience and evaluating long-term land surface elevation trends. Specifically, because shrink-swell movements seem to increase with increasing drought intensity, duration and occurrence. This study aimed to identify a universally applicable and straightforward method to predict shallow subsurface deformation based on changes in volumetric water content (VWC) or soil water suction and laboratory-measured shrinkage behaviour. An existing method based on changes in VWC and a newly proposed method based on changes in soil water suction were intercompared. The newly proposed method relates the vertical- or volumetric shrinkage to the matric suction, in combination with a depth-based overburden term. The input for both prediction models consisted of the shrinkage behaviour measured in a laboratory setup. The shrinkage behaviour was described as the change in sample- volume, height, the water content and soil water suction, during drying. Samples were collected from three sites in the Netherlands, where expansive clay minerals were identified and the models could be validated using extensometer data. Both prediction models proved to be able to capture seasonal shrinkage behaviour well. However, short-term (days to weeks) deformations were more challenging to predict due to the difficulty in calibrating VWC-sensors for the full VWC range. Moreover, seasonal swell was not predicted accurately for a site with a relatively low expansive clay mineral content. The Briaud model was the most accurate prediction model for shrinkage behaviour tested in this study.