Rapid non-destructive methods for estimating shear strength of sandy soils

Assessing soil mechanical properties at the field scale is crucial for preventing compaction from agricultural traffic. Conventional laboratory methods are destructive, labor-intensive, and require geostatistical methods to assess field variations. This study evaluates the potential of on-the-go proximal soil sensors (PSS), specifically electromagnetic induction (EMI) and gamma-ray spectroscopy (GRS) in combination with terrain attributes (TA) to estimate soil shear strength (τ). For this study, bulk soil samples and intact soil cores (100 cm³) were collected at 23 locations from depths of 0.15 m and 0.40 m across a 35.5 ha field, representing both the topsoil and subsoil layers. Multiple linear regression (MLR) models were developed to predict cohesion (c) and angle of internal friction (φ), with τ derived from these estimates. For comparison, a pedotransfer function (PTF) based on intrinsic soil properties and bulk density was also applied. Results showed that PSS–TA models achieved substantially higher accuracy (RMSE = 11 kPa at 0.15 m; 8 kPa at 0.40 m) than PTF (26 kPa and 27 kPa, respectively). While PTF remains a practical alternative when sensors are inaccessible, the findings demonstrate a clear advantage in integrating PSS and TA for the rapid, non-invasive, field-scale assessment of τ. • Novel method developed to estimate soil shear strength (τ). • Proximal soil sensors (PSS) and terrain attributes (TA) effectively estimated τ. • PSS+TA estimated τ more accurately than the pedotransfer function (PTF). • Subsoil cohesion (c) estimates showed higher accuracy. • Topsoil angle of internal friction (φ) estimates were more accurate.