Subsoil shear strength - Measurements and prediction models based on readily available soil properties

Traffic-induced long-term damage to agricultural subsoils is a serious threat in modern, mechanized agriculture. Soil failure due to shear stresses is often not considered in soil compaction models but probably contributes significantly to soil structure deterioration. We measured shear strength for a total of 720 undisturbed soil cores collected at nine different locations in Denmark. Soil clay content ranged from 0.025 to 0.375 kg kg(-1). We sampled at soil depths 0.3, 0.5 and 0.8m and drained the soil cores to either -50, -100 or -300 hPa matric potential prior to shear tests. We used a shear annulus device to apply shear stress to the soil cores. The normal load, NL, in tests was either 30, 60, 90, 120, 150 or 180 kPa. Soil shear strength, tau, was estimated as the peak (maximum) shear stress at soil failure. Soil cohesion and angle of internal friction was estimated from linear regression of tau and NL. Multiple regression indicated that soil cohesion was well predicted by soil organic matter, clay content, the initial (preload) suction stress, sigma(w) and soil bulk density. sigma (w) was calculated from water saturation and the matric potential. It was superior to quantity expressions of soil water (volumetric water content and water ratio) in explaining the trends in soil cohesion. Quantity of soil water should not be used in prediction models across soil types. The angle of internal friction correlated poorly to soil properties. Soil shear strength at a given NL could be well described by a model combining the above soil properties with the NL. This pedotransfer function predicted reasonably well the measured shear strength from two independent data sets. More studies are needed to evaluate a range of methodological aspects and for inclusion of more clay-holding soils. We encourage the inclusion of soil shear failure prediction in soil compaction models and suggest a specific procedure for this.